From f55c2b92676e4a7eb06efbc8d9a908a7de62541a Mon Sep 17 00:00:00 2001 From: morluto <76467478+morluto@users.noreply.github.com> Date: Tue, 29 Sep 2026 11:22:00 +0800 Subject: [PATCH] refactor(mcp): simplify tool inputs and inline results --- src/application/BrowserEvidence.ts | 4 +- src/application/BrowserScenarioEvidence.ts | 2 +- src/application/CapabilityInventory.ts | 1 - src/application/EnhancedTools.ts | 34 +-- src/browser/BrowserScenario.test.ts | 56 ++-- .../PlaywrightBrowserScenarioProvider.test.ts | 24 -- src/browser/PlaywrightScenarioArtifacts.ts | 1 - src/browser/PlaywrightScenarioEvents.test.ts | 16 - src/cliEvidenceCommands.ts | 8 +- .../applicationWorkflowInputContracts.ts | 7 - src/contracts/browserScenarioToolContracts.ts | 55 +--- src/contracts/browserToolContracts.ts | 35 +-- src/contracts/electronToolContracts.ts | 11 - src/contracts/enhancedInputs.ts | 30 +- .../functionWorkflowToolContracts.ts | 4 +- ...vascriptRuntimeObservationToolContracts.ts | 9 - src/contracts/managedWorkflowToolContracts.ts | 6 - src/contracts/promptContracts.ts | 4 +- src/contracts/sessionToolSchemas.ts | 4 - src/contracts/toolContractExamples.ts | 1 + src/contracts/toolContracts.test.ts | 2 +- src/contracts/toolContracts.ts | 34 +-- src/contracts/toolEffects.ts | 1 - src/contracts/toolOutputSchemaGroups.ts | 9 +- src/contracts/toolSchemaMetadata.ts | 10 +- src/domain/artifactComparison.ts | 4 - src/domain/browserScenario.ts | 32 +- src/domain/browserScenarioValues.ts | 284 ++++++++++-------- src/domain/symbolAnalysis.ts | 24 +- src/domain/webContentArtifact.ts | 11 - src/domain/webScreenshot.ts | 8 +- src/server/ProcessCaptureElicitation.ts | 3 - src/server/createServer.ts | 2 +- src/server/recordDerivedEvidence.ts | 2 +- .../compareExportShapes.ts | 4 +- .../compareSourceToBundle.ts | 3 +- .../compareVersions.ts | 4 +- src/server/registerArtifactComparisonTool.ts | 6 +- src/server/registerBrowserScenarioTool.ts | 4 +- src/server/registerEnhancedTools.ts | 14 - src/server/registerFunctionComparisonTool.ts | 5 +- src/server/registerInvestigationTools.ts | 31 +- .../registerInvestigationTools/helpers.ts | 5 +- .../registerInvestigationTools/types.ts | 2 +- .../compareManagedMembers.ts | 43 ++- .../verifyManagedNativeBoundaries.ts | 53 ++-- src/server/registerOfficialTools.ts | 69 +---- src/server/registerProcessComparisonTool.ts | 7 +- src/server/registerSessionTools.ts | 1 - .../applicationEvidenceMcp.test.ts | 1 - .../applicationExportShapeMcp.test.ts | 2 +- ...cdpBrowserProvider.document_script.test.ts | 4 +- tests/boundary/cli/browserCli.test.ts | 2 +- tests/boundary/mcp/artifactMcp.test.ts | 1 - tests/boundary/mcp/browserMcp.test.ts | 2 +- tests/boundary/mcp/browserScenarioMcp.test.ts | 72 ++--- tests/boundary/mcp/electronMcp.test.ts | 2 +- ....01-lists-the-complete-target-open.test.ts | 34 ++- ...3-returns-typed-graph-failures-and.test.ts | 9 +- ...4-uses-complete-inventories-inline.test.ts | 2 +- ...ts-incomplete-function-dossiers-at.test.ts | 26 -- .../mcp/enhancedTools.inputValidation.test.ts | 2 +- .../mcp/functionComparisonMcp.test.ts | 3 +- tests/boundary/mcp/ghidraRoutingMcp.test.ts | 1 - .../mcp/investigationAnalysisMcp.test.ts | 5 - .../mcp/processTraceAdaptersParity.test.ts | 1 - tests/boundary/mcp/serverIdentity.test.ts | 2 - tests/boundary/mcp/serverUnknownsMcp.test.ts | 20 +- tests/boundary/mcp/serverWorkflowMcp.test.ts | 26 ++ tests/boundary/mcp/sessionMcp.test.ts | 3 +- 70 files changed, 468 insertions(+), 741 deletions(-) diff --git a/src/application/BrowserEvidence.ts b/src/application/BrowserEvidence.ts index 52ebd02b58..c9728fcecf 100644 --- a/src/application/BrowserEvidence.ts +++ b/src/application/BrowserEvidence.ts @@ -124,8 +124,8 @@ const browserParameters = ( }; if (isScreenshotComparison(input)) return { - before_artifact_uri: input.before.uri, - after_artifact_uri: input.after.uri, + before_artifact_sha256: input.before.sha256, + after_artifact_sha256: input.after.sha256, channel_threshold: input.channel_threshold, maximum_pixels: input.maximum_pixels, }; diff --git a/src/application/BrowserScenarioEvidence.ts b/src/application/BrowserScenarioEvidence.ts index 94a3d3ebb2..733a1f5ddf 100644 --- a/src/application/BrowserScenarioEvidence.ts +++ b/src/application/BrowserScenarioEvidence.ts @@ -30,7 +30,7 @@ const browserScenarioParameters = ( }), ), capture: scenario.capture, - limits: scenario.limits, + limits: { ...scenario.limits }, }); /** Create Evidence without retaining resolved scenario secret values. */ diff --git a/src/application/CapabilityInventory.ts b/src/application/CapabilityInventory.ts index 9dfb8988fd..8d5b9bebb3 100644 --- a/src/application/CapabilityInventory.ts +++ b/src/application/CapabilityInventory.ts @@ -72,7 +72,6 @@ type AvailabilityContext = { }; const ENHANCED_REQUIREMENTS: Readonly> = { - swift_classes: ["list_procedures"], get_objc_classes: ["list_names"], get_objc_protocols: ["list_names"], batch_decompile: ["procedure_pseudo_code"], diff --git a/src/application/EnhancedTools.ts b/src/application/EnhancedTools.ts index 31d38d3b1b..f188a516f3 100644 --- a/src/application/EnhancedTools.ts +++ b/src/application/EnhancedTools.ts @@ -26,7 +26,6 @@ import { categorizeSwiftTypes, discoverObjcClasses, discoverObjcProtocols, - discoverSwiftClasses, } from "../domain/symbolAnalysis.js"; import { jsonValueSchema, type JsonValue } from "../domain/jsonValue.js"; @@ -63,12 +62,6 @@ export class EnhancedTools { if (name === "analyze_function" || name === "inspect_native_api") return this.#executeFunctionAnalysis(name, input, signal); switch (name) { - case "swift_classes": { - const parsed = enhancedInputSchemas.swift_classes.safeParse(input); - return parsed.success - ? this.executeValidated({ name, input: parsed.data }, signal) - : invalidEnhancedInput(name, parsed.error); - } case "get_objc_classes": { const parsed = enhancedInputSchemas.get_objc_classes.safeParse(input); return parsed.success @@ -89,8 +82,13 @@ export class EnhancedTools { ? this.executeValidated({ name, input: parsed.data }, signal) : invalidEnhancedInput(name, parsed.error); } - case "analyze_swift_types": - return this.executeValidated({ name, input: {} }, signal); + case "analyze_swift_types": { + const parsed = + enhancedInputSchemas.analyze_swift_types.safeParse(input); + return parsed.success + ? this.executeValidated({ name, input: parsed.data }, signal) + : invalidEnhancedInput(name, parsed.error); + } case "find_xrefs_to_name": { const parsed = enhancedInputSchemas.find_xrefs_to_name.safeParse(input); return parsed.success @@ -112,8 +110,6 @@ export class EnhancedTools { signal?: AbortSignal, ): EnhancedResult { switch (call.name) { - case "swift_classes": - return this.#swiftClasses(call.input.pattern, signal); case "get_objc_classes": return this.#objcClasses(call.input.pattern, signal); case "get_objc_protocols": @@ -123,7 +119,7 @@ export class EnhancedTools { case "get_call_graph": return this.#callGraph(call.input, signal); case "analyze_swift_types": - return this.#analyzeSwiftTypes(signal); + return this.#analyzeSwiftTypes(call.input, signal); case "find_xrefs_to_name": return this.#findXrefs(call.input.name, signal); case "binary_overview": @@ -216,13 +212,6 @@ export class EnhancedTools { return ok(jsonValueSchema.parse(projectNativeApiInspection(dossier))); } - async #swiftClasses(pattern: string, signal?: AbortSignal): EnhancedResult { - const procedures = await this.#allAddressed("list_procedures", signal); - return procedures.ok - ? ok(discoverSwiftClasses(procedures.value, pattern)) - : procedures; - } - async #objcClasses(pattern: string, signal?: AbortSignal): EnhancedResult { const names = await this.#allAddressed("list_names", signal); return names.ok ? ok(discoverObjcClasses(names.value, pattern)) : names; @@ -333,10 +322,13 @@ export class EnhancedTools { return ok(graph); } - async #analyzeSwiftTypes(signal?: AbortSignal): EnhancedResult { + async #analyzeSwiftTypes( + input: z.output, + signal?: AbortSignal, + ): EnhancedResult { const procedures = await this.#allAddressed("list_procedures", signal); return procedures.ok - ? ok(categorizeSwiftTypes(procedures.value)) + ? ok(categorizeSwiftTypes(procedures.value, input)) : procedures; } diff --git a/src/browser/BrowserScenario.test.ts b/src/browser/BrowserScenario.test.ts index 77174ef0ff..9de9a35f11 100644 --- a/src/browser/BrowserScenario.test.ts +++ b/src/browser/BrowserScenario.test.ts @@ -2,6 +2,7 @@ import { describe, expect, it } from "vitest"; import { BrowserScenarioSecrets } from "./BrowserScenarioSecrets.js"; import { browserScenarioSchema } from "../domain/browserScenario.js"; +import { BROWSER_SCENARIO_CAPTURE_LIMITS } from "../domain/browserScenarioValues.js"; const secret = (secret_id: string) => ({ source: "secret", secret_id }); const literal = (value: string) => ({ @@ -23,14 +24,6 @@ const baseScenario = () => ({ query: [{ name: "token", value: secret("session") }], }, allowed_origins: ["https://api.example.test", "https://app.example.test"], - environment: { - viewport: { width: 1_280, height: 720 }, - locale: "en-US", - timezone: "UTC", - color_scheme: "dark", - reduced_motion: "reduce", - service_workers: "block", - }, actions: [ { step_id: "login", @@ -124,22 +117,6 @@ const baseScenario = () => ({ at_end: ["dom", "storage"], events: ["console", "page-errors", "network", "websockets"], }, - limits: { - max_duration_ms: 60_000, - action_timeout_ms: 5_000, - navigation_timeout_ms: 10_000, - max_events: 2_000, - max_frames: 100, - max_workers: 20, - max_popups: 10, - max_websockets: 100, - max_dom_nodes: 10_000, - max_accessibility_nodes: 10_000, - max_screenshots: 16, - max_screenshot_bytes: 4 * 1_024 * 1_024, - max_storage_entries: 256, - max_total_metadata_bytes: 4 * 1_024 * 1_024, - }, }); describe("browserScenarioSchema", () => { @@ -149,13 +126,40 @@ describe("browserScenarioSchema", () => { "https://api.example.test", "https://app.example.test", ]); - expect(parsed.environment.viewport.device_scale_factor).toBe(1); + expect(parsed.environment).toMatchObject({ + viewport: { width: 1_280, height: 720, device_scale_factor: 1 }, + locale: "en-US", + timezone: "UTC", + color_scheme: "light", + reduced_motion: "reduce", + service_workers: "block", + }); + expect(parsed.request_replay).toEqual({ + mode: "exact", + unmatched: "abort", + routes: expect.any(Array), + }); + expect(parsed.capture).toEqual({ + after_each_step: ["screenshot", "url", "accessibility"], + at_end: ["dom", "storage"], + events: ["console", "page-errors", "network", "websockets"], + }); expect(parsed.redaction.header_names).toEqual([ "authorization", "cookie", "proxy-authorization", "set-cookie", ]); + expect(parsed.limits).toEqual(BROWSER_SCENARIO_CAPTURE_LIMITS); + }); + + it("rejects caller supplied capture limits", () => { + expect( + browserScenarioSchema.safeParse({ + ...baseScenario(), + limits: { max_duration_ms: 1 }, + }).success, + ).toBe(false); }); it("redacts overlapping secrets longest-first", () => { @@ -205,7 +209,9 @@ describe("browserScenarioSchema", () => { expect(browserScenarioSchema.safeParse(scenario).success).toBe(false); } }); +}); +describe("browserScenarioSchema credential handling", () => { it("rejects redirects outside declared origins", () => { const scenario = baseScenario(); scenario.request_replay.routes[0]!.response = { diff --git a/src/browser/PlaywrightBrowserScenarioProvider.test.ts b/src/browser/PlaywrightBrowserScenarioProvider.test.ts index 41da6d79c3..1cfd26ec97 100644 --- a/src/browser/PlaywrightBrowserScenarioProvider.test.ts +++ b/src/browser/PlaywrightBrowserScenarioProvider.test.ts @@ -58,14 +58,6 @@ const scenario = ( }, start_url: { url: "https://app.example.test/" }, allowed_origins: ["https://app.example.test"], - environment: { - viewport: { width: 1_280, height: 720 }, - locale: "en-US", - timezone: "UTC", - color_scheme: "light", - reduced_motion: "reduce", - service_workers: "block", - }, actions: Array.from({ length: options.actions ?? 1 }, (_, index) => ({ step_id: `wait_${index}`, action: "wait_for_timeout", @@ -83,22 +75,6 @@ const scenario = ( at_end: [], events: options.events ?? [], }, - limits: { - max_duration_ms: 10_000, - action_timeout_ms: 1_000, - navigation_timeout_ms: 1_000, - max_events: 100, - max_frames: 10, - max_workers: 10, - max_popups: 10, - max_websockets: 10, - max_dom_nodes: 100, - max_accessibility_nodes: 100, - max_screenshots: 2, - max_screenshot_bytes: 1_024, - max_storage_entries: 10, - max_total_metadata_bytes: 16_384, - }, }); type SnapshotKind = BrowserScenario["capture"]["after_each_step"][number]; diff --git a/src/browser/PlaywrightScenarioArtifacts.ts b/src/browser/PlaywrightScenarioArtifacts.ts index 17bdd06479..f75145b874 100644 --- a/src/browser/PlaywrightScenarioArtifacts.ts +++ b/src/browser/PlaywrightScenarioArtifacts.ts @@ -122,7 +122,6 @@ const captureScreenshot = async ( return { state: "captured" as const, value: { - uri: `rea://web-screenshot/sha256/${sha256}`, sha256, bytes: bytes.byteLength, media_type: "image/png" as const, diff --git a/src/browser/PlaywrightScenarioEvents.test.ts b/src/browser/PlaywrightScenarioEvents.test.ts index e0084b6966..ee7e3d1208 100644 --- a/src/browser/PlaywrightScenarioEvents.test.ts +++ b/src/browser/PlaywrightScenarioEvents.test.ts @@ -52,22 +52,6 @@ describe("PlaywrightScenarioEvents", () => { at_end: [], events: ["page-errors"], }, - limits: { - max_duration_ms: 10_000, - action_timeout_ms: 1_000, - navigation_timeout_ms: 1_000, - max_events: 10, - max_frames: 10, - max_workers: 10, - max_popups: 10, - max_websockets: 10, - max_dom_nodes: 100, - max_accessibility_nodes: 100, - max_screenshots: 1, - max_screenshot_bytes: 1_024, - max_storage_entries: 10, - max_total_metadata_bytes: 1_000_000, - }, }); const secrets = BrowserScenarioSecrets.resolve(scenario, {}); if (secrets === undefined) throw new Error("Expected resolved secrets"); diff --git a/src/cliEvidenceCommands.ts b/src/cliEvidenceCommands.ts index 4d3b0e6955..0e4bcf1859 100644 --- a/src/cliEvidenceCommands.ts +++ b/src/cliEvidenceCommands.ts @@ -19,13 +19,13 @@ import { import { loadConfiguredPermissionAuthority } from "./application/PermissionConfiguration.js"; import type { AppConfig } from "./config.js"; -/** Register filesystem-gated Evidence v2 commands. */ +/** Register filesystem-gated Evidence commands. */ export const registerEvidenceCommands = ( cli: ReturnType, logger: Logger, ): void => { cli.command(CLI_COMMANDS.evidenceImport, { - description: "Validate and import a bounded local Evidence v2 bundle", + description: "Validate and import a bounded local Evidence bundle", args: z.object({ path: z.string().describe("Evidence bundle JSON path"), }), @@ -45,7 +45,7 @@ export const registerEvidenceCommands = ( }), }); cli.command(CLI_COMMANDS.evidenceExport, { - description: "Validate and atomically export canonical Evidence v2 JSON", + description: "Validate and atomically export canonical Evidence JSON", args: z.object({ source: z.string().describe("Existing evidence bundle JSON path"), output: z.string().describe("Canonical output JSON path"), @@ -73,7 +73,7 @@ export const registerEvidenceCommands = ( }); cli.command(CLI_COMMANDS.compare, { aliases: ["compare-bundles"], - description: "Compare two canonical Evidence v2 bundles", + description: "Compare two canonical Evidence bundles", args: z.object({ left: z.string().describe("Left Evidence bundle JSON path"), right: z.string().describe("Right Evidence bundle JSON path"), diff --git a/src/contracts/applicationWorkflowInputContracts.ts b/src/contracts/applicationWorkflowInputContracts.ts index 1a33e4e0e0..4ad6462131 100644 --- a/src/contracts/applicationWorkflowInputContracts.ts +++ b/src/contracts/applicationWorkflowInputContracts.ts @@ -51,8 +51,6 @@ const compareApplicationVersionsFacts = { right_native_observations: compareApplicationVersionsInputSchema.shape.right_native_observations, right_native_observation_evidence_ids: z.array(evidenceIdSchema).default([]), - unknown_registry_approved: - compareApplicationVersionsInputSchema.shape.unknown_registry_approved, } as const; const applicationEvidenceReferenceSchema = z.union([ @@ -95,9 +93,6 @@ export const compareApplicationVersionsRequestSchema = z const compareSourceToBundleFacts = { reference: compareSourceToBundleInputSchema.shape.reference, - limits: compareSourceToBundleInputSchema.shape.limits, - unknown_registry_approved: - compareSourceToBundleInputSchema.shape.unknown_registry_approved, } as const; /** Historical-source comparison accepting full application Evidence or a ledger reference. */ @@ -121,8 +116,6 @@ const compareJavaScriptExportShapesFacts = { compareJavaScriptExportShapesInputSchema.shape.right_module_path, right_export_name: compareJavaScriptExportShapesInputSchema.shape.right_export_name, - unknown_registry_approved: - compareJavaScriptExportShapesInputSchema.shape.unknown_registry_approved, } as const; /** MCP/CLI export-shape request accepting full Evidence or ledger references. */ diff --git a/src/contracts/browserScenarioToolContracts.ts b/src/contracts/browserScenarioToolContracts.ts index a732607af8..82e103bc9c 100644 --- a/src/contracts/browserScenarioToolContracts.ts +++ b/src/contracts/browserScenarioToolContracts.ts @@ -1,7 +1,7 @@ import type { ToolContract } from "./toolContracts.js"; import { toolContractMetadata } from "./toolEffects.js"; import { evidenceResultOf } from "./toolOutputSchemas.js"; -import { browserScenarioSchema } from "../domain/browserScenario.js"; +import { browserScenarioInputSchema } from "../domain/browserScenario.js"; import { browserScenarioCaptureSchema } from "../domain/browserScenarioCapture.js"; import type { JsonValue } from "../domain/jsonValue.js"; @@ -15,56 +15,13 @@ const example: Record = { }, start_url: { url: "https://app.example.test/", query: [] }, allowed_origins: ["https://app.example.test"], - environment: { - viewport: { width: 1_280, height: 720, device_scale_factor: 1 }, - locale: "en-US", - timezone: "UTC", - color_scheme: "light", - reduced_motion: "reduce", - service_workers: "block", - }, actions: [ { - step_id: "ready", - action: "wait_for", - locator: { kind: "test_id", value: "application-ready" }, - state: "visible", + step_id: "settle", + action: "wait_for_timeout", + duration_ms: 250, }, ], - storage: { cookies: [], local_storage: [], session_storage: [] }, - request_replay: { mode: "disabled" }, - secrets: [], - redaction: { - secret_values: "replace-with-secret-reference", - query_parameter_names: [], - header_names: [ - "authorization", - "cookie", - "proxy-authorization", - "set-cookie", - ], - }, - capture: { - after_each_step: ["screenshot", "url", "accessibility"], - at_end: ["dom", "history", "storage"], - events: ["console", "page-errors", "network", "websockets"], - }, - limits: { - max_duration_ms: 60_000, - action_timeout_ms: 5_000, - navigation_timeout_ms: 10_000, - max_events: 2_000, - max_frames: 100, - max_workers: 20, - max_popups: 10, - max_websockets: 100, - max_dom_nodes: 10_000, - max_accessibility_nodes: 10_000, - max_screenshots: 16, - max_screenshot_bytes: 4_194_304, - max_storage_entries: 256, - max_total_metadata_bytes: 4_194_304, - }, }; /** Controlled browser scenario contract shared by MCP and catalog generation. */ @@ -73,9 +30,9 @@ export const BROWSER_SCENARIO_TOOL_CONTRACTS = [ name: "capture_browser_scenario", ...toolContractMetadata("capture_browser_scenario"), description: - "Run one approved, bounded browser scenario through Playwright. Launch mode owns and terminates a temporary browser profile; connect mode attaches to one exact loopback CDP target and disconnects without terminating the external browser. Only the fixed action vocabulary is accepted. Navigation, storage, requests, redirects, secrets, redaction, captures, and limits must be declared before execution. Returns step-indexed Evidence with screenshots, DOM, accessibility, URL/history/storage state, runtime events, explicit missing/truncated sections, and equality eligibility.", + "Run a controlled browser scenario when passive observation cannot exercise the application. Choose launch mode with an approved executable or connect mode with an approved loopback CDP target, declare the exact start origin scope, and provide explicit actions. Environment, empty storage, disabled request replay, redaction, and a URL-only final capture have safe defaults; declare secrets, initial storage, exact request replay, or additional retained artifacts only when needed. Provider-owned duration and collection budgets are fixed. Returns step-indexed Evidence with explicit missing/truncated sections and equality eligibility.", kind: "browser-provider", - inputSchema: browserScenarioSchema, + inputSchema: browserScenarioInputSchema, outputSchema: evidenceResultOf(browserScenarioCaptureSchema), examples: [ { title: "Capture a deterministic browser scenario", input: example }, diff --git a/src/contracts/browserToolContracts.ts b/src/contracts/browserToolContracts.ts index 1550cce920..e9b896c4fc 100644 --- a/src/contracts/browserToolContracts.ts +++ b/src/contracts/browserToolContracts.ts @@ -138,7 +138,7 @@ export const BROWSER_TOOL_CONTRACTS = [ name: "inspect_web_page", ...toolContractMetadata("inspect_web_page"), description: - "Passively inspect one approved page target through CDP without evaluating JavaScript, navigating, clicking, closing, or mutating the page. Returns bounded DOM structure, accessibility, scripts, resources, attach-window network and console metadata, workers, and redacted storage inventory as Evidence v2.", + "Passively inspect one approved page target through CDP without evaluating JavaScript, navigating, clicking, closing, or mutating the page. Returns bounded DOM structure, accessibility, scripts, resources, attach-window network and console metadata, workers, and redacted storage inventory as Evidence.", kind: "browser-provider", inputSchema: inspectWebPageInputSchema, outputSchema: inspectionOutputSchema, @@ -149,38 +149,6 @@ export const BROWSER_TOOL_CONTRACTS = [ cdp_endpoint: endpoint, allowed_origins: [origin], target_id: "TARGET_ID_FROM_LIST_BROWSER_TARGETS", - observation_ms: 500, - include_accessibility_text: false, - include_console_text: false, - include_json_body_shapes: false, - include_websocket_shapes: false, - include_script_sources: false, - include_storage_keys: false, - include_storage_fingerprints: false, - limits: { - max_frames: 200, - max_dom_nodes: 2_000, - max_ax_nodes: 2_000, - max_ax_text_field_bytes: 1_024, - max_total_ax_text_bytes: 65_536, - max_scripts: 200, - max_resources: 2_000, - max_workers: 500, - max_storage_keys: 1_000, - max_script_source_bytes: 1_048_576, - max_total_script_source_bytes: 4_194_304, - max_network_events: 1_000, - max_console_events: 200, - max_console_text_field_bytes: 1_024, - max_total_console_text_bytes: 65_536, - max_json_body_bytes: 1_048_576, - max_total_json_body_bytes: 4_194_304, - max_json_shape_nodes: 5_000, - max_json_shape_depth: 20, - max_websocket_events: 500, - max_websocket_shape_bytes: 65_536, - max_total_websocket_shape_bytes: 1_048_576, - }, }, }, ], @@ -327,7 +295,6 @@ export const BROWSER_TOOL_CONTRACTS = [ function exampleScreenshot() { return { - uri: "rea://web-screenshot/sha256/153cf6c9a526a63053a37b10234c2fd85df38887c2dc0a800d90abfa6631d01c", sha256: "153cf6c9a526a63053a37b10234c2fd85df38887c2dc0a800d90abfa6631d01c", bytes: 70, media_type: "image/png", diff --git a/src/contracts/electronToolContracts.ts b/src/contracts/electronToolContracts.ts index 22370e7a1d..dd8bd0ae3b 100644 --- a/src/contracts/electronToolContracts.ts +++ b/src/contracts/electronToolContracts.ts @@ -83,17 +83,6 @@ export const ELECTRON_TOOL_CONTRACTS = [ cdp_endpoint: endpoint, allowed_file_roots: [root], target_id: "TARGET_ID_FROM_LIST_ELECTRON_TARGETS", - observation_ms: 100, - include_script_sources: false, - limits: { - max_frames: 200, - max_dom_nodes: 2_000, - max_scripts: 500, - max_resources: 2_000, - max_workers: 500, - max_script_source_bytes: 1_048_576, - max_total_script_source_bytes: 4_194_304, - }, }, }, ], diff --git a/src/contracts/enhancedInputs.ts b/src/contracts/enhancedInputs.ts index 889b94fcc6..63f9550575 100644 --- a/src/contracts/enhancedInputs.ts +++ b/src/contracts/enhancedInputs.ts @@ -3,15 +3,10 @@ import { z } from "zod"; const traceLiteralInputSchema = z.strictObject({ query: z.string().min(1), case_sensitive: z.boolean().default(false), - unknown_registry_approved: z - .literal(true) - .optional() - .describe("Explicit approval to record bounded residuals durably"), }); /** Input schemas shared by MCP registration and enhanced application dispatch. */ export const enhancedInputSchemas = { - swift_classes: z.strictObject({ pattern: z.string().default("") }), get_objc_classes: z.strictObject({ pattern: z.string().default("") }), get_objc_protocols: z.strictObject({}), batch_decompile: z.strictObject({ @@ -23,7 +18,18 @@ export const enhancedInputSchemas = { address: z.string().describe("A provider-normalized procedure address"), direction: z.enum(["forward", "backward"]).default("forward"), }), - analyze_swift_types: z.strictObject({}), + analyze_swift_types: z.strictObject({ + category: z + .enum(["classes", "structs", "enums", "protocols", "extensions", "other"]) + .optional() + .describe("Limit results to one Swift symbol category."), + pattern: z + .string() + .optional() + .describe( + "Case-sensitive literal filter applied to mangled symbol names.", + ), + }), find_xrefs_to_name: z.strictObject({ name: z.string() }), binary_overview: z.strictObject({}), analyze_function: z.strictObject({ @@ -31,12 +37,6 @@ export const enhancedInputSchemas = { }), inspect_native_api: z.strictObject({ procedure: z.string().describe("A procedure name or address"), - unknown_registry_approved: z - .literal(true) - .optional() - .describe( - "Explicit approval to record unsupported native API branches durably", - ), }), trace_feature: traceLiteralInputSchema, find_code_for_string: traceLiteralInputSchema, @@ -47,12 +47,6 @@ export const enhancedInputSchemas = { .describe("An optional provider-normalized destination address") .optional(), direction: z.enum(["forward", "backward"]).default("forward"), - unknown_registry_approved: z - .literal(true) - .optional() - .describe( - "Explicit approval to record unresolved provider results durably", - ), }), } as const; diff --git a/src/contracts/functionWorkflowToolContracts.ts b/src/contracts/functionWorkflowToolContracts.ts index 3816f8a20e..45125faede 100644 --- a/src/contracts/functionWorkflowToolContracts.ts +++ b/src/contracts/functionWorkflowToolContracts.ts @@ -36,10 +36,10 @@ const functionWorkflow = ( export const FUNCTION_WORKFLOW_TOOL_CONTRACTS = [ functionWorkflow( "analyze_function", - "Return one complete function dossier inline, including identity, provider-specific pseudocode and assembly, comments, calls, references, referenced strings/names, local CFG blocks, and any available native API boundary observations.", + "Use for a specific native function when you need its behavior and surrounding evidence in one call. Pass a procedure symbol/name or provider-returned address, typically selected from search_procedures or list_procedures. Returns identity, provider-specific pseudocode and assembly, comments, calls, references, referenced strings/names, local CFG blocks, and available native API boundary observations. For only one facet, use the focused procedure tool; for structured native API reconstruction, use inspect_native_api.", ), functionWorkflow( "inspect_native_api", - "Reconstruct one native function boundary with structured return/parameter types, confidence, evidence, jump-table dispatch/data/target mappings, unsupported branches, and residual unknowns.", + "Use after identifying a function that calls or implements a native API boundary when you need structured return/parameter types and dispatch evidence. Pass its procedure symbol/name or provider-returned address. Returns confidence, evidence, jump-table dispatch/data/target mappings, unsupported branches, and residual unknowns; use analyze_function for a general function dossier.", ), ] as const satisfies readonly ToolContract[]; diff --git a/src/contracts/javascriptRuntimeObservationToolContracts.ts b/src/contracts/javascriptRuntimeObservationToolContracts.ts index 9d44419e0e..d14c3203ea 100644 --- a/src/contracts/javascriptRuntimeObservationToolContracts.ts +++ b/src/contracts/javascriptRuntimeObservationToolContracts.ts @@ -46,16 +46,7 @@ export const JAVASCRIPT_RUNTIME_OBSERVATION_TOOL_CONTRACTS = [ input: { inspector_endpoint: endpoint, allowed_file_roots: [root], - allowed_origins: [], target_id: "TARGET_ID_FROM_LIST_JAVASCRIPT_RUNTIME_TARGETS", - observation_ms: 100, - limits: { - max_events: 10_000, - max_scripts: 2_000, - max_execution_contexts: 1_000, - max_location_bytes: 16_384, - max_total_metadata_bytes: 4_194_304, - }, }, }, ], diff --git a/src/contracts/managedWorkflowToolContracts.ts b/src/contracts/managedWorkflowToolContracts.ts index 7e05987388..1733fd9078 100644 --- a/src/contracts/managedWorkflowToolContracts.ts +++ b/src/contracts/managedWorkflowToolContracts.ts @@ -1,7 +1,5 @@ import { z } from "zod"; -import { compareManagedMembersInputSchema } from "../domain/managedMemberComparison.js"; -import { managedNativeVerificationInputSchema } from "../domain/managedNativeVerification.js"; import { managedReconstructionImportInputSchema } from "../domain/managedReconstruction.js"; import { managedRuntimeCorrelationInputSchema } from "../domain/managedRuntimeCorrelation.js"; import type { ToolContract } from "./toolContracts.js"; @@ -35,8 +33,6 @@ export const compareManagedMembersReferenceInputSchema = z .strictObject({ left_evidence_id: managedEvidenceIdSchema, right_evidence_id: managedEvidenceIdSchema, - unknown_registry_approved: - compareManagedMembersInputSchema.shape.unknown_registry_approved, }) .superRefine((input, context) => { if (input.left_evidence_id === input.right_evidence_id) @@ -67,8 +63,6 @@ export const managedNativeVerificationReferenceInputSchema = z .array(nativeObservationEvidenceIdSchema) .min(1) .describe("Unique session-owned native observation Evidence IDs"), - unknown_registry_approved: - managedNativeVerificationInputSchema.shape.unknown_registry_approved, }) .superRefine((input, context) => { const ids = new Set(); diff --git a/src/contracts/promptContracts.ts b/src/contracts/promptContracts.ts index e55611929c..5020ad2eb2 100644 --- a/src/contracts/promptContracts.ts +++ b/src/contracts/promptContracts.ts @@ -223,7 +223,7 @@ export const PROMPT_CONTRACTS = [ name: "verify_reconstruction", title: "Verify a reconstruction", description: - "Evaluate a finite reconstruction specification against retained Evidence v2 comparisons without broadening pass results into global equivalence claims.", + "Evaluate a finite reconstruction specification against retained Evidence comparisons without broadening pass results into global equivalence claims.", objective: "Produce per-claim pass, fail, or unknown results backed by compatible comparison Evidence and exact authority requirements.", arguments: { @@ -416,7 +416,7 @@ export const PROMPT_CONTRACTS = [ { tools: ["record_unknown"], instruction: - "Treat sampling gaps, truncated output, external behavior, and cleanup uncertainty as residual unknowns; persist them only with separate explicit approval.", + "Treat sampling gaps, truncated output, external behavior, and cleanup uncertainty as residual unknowns; support them with relevant Evidence IDs and focused recommended probes.", }, ], }, diff --git a/src/contracts/sessionToolSchemas.ts b/src/contracts/sessionToolSchemas.ts index 0ba2a4918b..56b2d94422 100644 --- a/src/contracts/sessionToolSchemas.ts +++ b/src/contracts/sessionToolSchemas.ts @@ -44,10 +44,6 @@ export const processComparisonInputSchema = z.strictObject({ right_evidence_id: z.string().regex(/^ev_[a-f0-9]{64}$/u), trace_spec: processTraceSpecificationSchema.optional(), max_capture_age_ms: z.number().int().nonnegative().optional(), - unknown_registry_approved: z - .literal(true) - .optional() - .describe("Explicit approval to record capture disagreement durably"), }); /** Residual-unknown list filters. */ diff --git a/src/contracts/toolContractExamples.ts b/src/contracts/toolContractExamples.ts index 548b470eab..38122e770c 100644 --- a/src/contracts/toolContractExamples.ts +++ b/src/contracts/toolContractExamples.ts @@ -35,6 +35,7 @@ export const TOOL_EXAMPLE_OVERRIDES: Readonly< get_call_graph: { address: "0x1000" }, find_xrefs_to_name: { name: "malloc" }, analyze_function: { procedure: "main" }, + analyze_swift_types: { category: "classes", pattern: "Account" }, inspect_native_api: { procedure: "main" }, trace_feature: { query: "license" }, find_code_for_string: { query: "authorization failed" }, diff --git a/src/contracts/toolContracts.test.ts b/src/contracts/toolContracts.test.ts index 6370e187ac..03f93b3caf 100644 --- a/src/contracts/toolContracts.test.ts +++ b/src/contracts/toolContracts.test.ts @@ -16,7 +16,7 @@ import { APPLICATION_TOOL_CONTRACTS } from "./applicationToolContracts.js"; describe("tool contract inventory", () => { it("publishes the canonical analysis and session tool inventory", () => { expect(OFFICIAL_TOOL_CONTRACTS).toHaveLength(36); - expect(ENHANCED_TOOL_CONTRACTS).toHaveLength(13); + expect(ENHANCED_TOOL_CONTRACTS).toHaveLength(12); expect(NATIVE_TOOL_CONTRACTS).toHaveLength(5); expect(ARTIFACT_TOOL_CONTRACTS.map(({ name }) => name)).toEqual([ "inspect_artifact", diff --git a/src/contracts/toolContracts.ts b/src/contracts/toolContracts.ts index 75a9e7abac..55c32ea8be 100644 --- a/src/contracts/toolContracts.ts +++ b/src/contracts/toolContracts.ts @@ -68,25 +68,16 @@ const official = ( name: Name, description: string, inputSchema: Schema, -) => { - const trackedInputSchema = inputSchema.extend({ - unknown_registry_approved: z - .literal(true) - .optional() - .describe( - "Explicit approval to record typed capability unavailability as a residual unknown", - ), - }); - return { +) => + ({ name, ...toolContractMetadata(name), description, kind: "official-proxy", - inputSchema: trackedInputSchema, + inputSchema, outputSchema: requireOutputSchema(officialOutputSchemas, name), - examples: examplesFor(name, trackedInputSchema), - } satisfies ToolContract; -}; + examples: examplesFor(name, inputSchema), + }) satisfies ToolContract; const enhanced = ( name: Name, @@ -304,11 +295,6 @@ export type OfficialToolName = (typeof OFFICIAL_TOOL_CONTRACTS)[number]["name"]; /** Bounded workflows composed from one or more bridge operations. */ export const ENHANCED_TOOL_CONTRACTS = [ - enhanced( - "swift_classes", - "Discover legacy-mangled Swift class procedures by scanning analyzed procedures; returns every matching entry. Use analyze_swift_types for other Swift kinds.", - enhancedInputSchemas.swift_classes, - ), enhanced( "get_objc_classes", "Discover and deduplicate every Objective-C class label, optionally filtering by literal substring; inspect matching metadata and references next.", @@ -331,7 +317,7 @@ export const ENHANCED_TOOL_CONTRACTS = [ ), enhanced( "analyze_swift_types", - "Categorize all analyzed procedure names into Swift classes, structs, enums, protocols, extensions, and other symbols.", + 'Categorize analyzed procedure names into Swift classes, structs, enums, protocols, extensions, and other symbols. Returns deduplicated names grouped with counts. Optionally select one category and/or apply a case-sensitive literal name filter; for example, use {category: "classes", pattern: "Account"} to find matching class symbols.', enhancedInputSchemas.analyze_swift_types, ), enhanced( @@ -341,13 +327,13 @@ export const ENHANCED_TOOL_CONTRACTS = [ ), enhanced( "binary_overview", - "Use immediately after opening a target to summarize the document, every segment with its length, and exhaustive procedure/string counts.", + "Use after opening a native binary for target metadata, every segment with its length, and exhaustive procedure/string counts. To choose a function, follow with search_procedures or list_procedures, then pass a returned name or address to analyze_function for a complete dossier.", enhancedInputSchemas.binary_overview, ), ...FUNCTION_WORKFLOW_TOOL_CONTRACTS, enhanced( "trace_feature", - "Trace a literal feature query through every matching string and procedure, their xrefs, and truthful containing-procedure resolution. Returns observations, operation count, and residual unknowns; unknown_registry_approved: true records residual unknowns without inferring reference kinds.", + "Trace a literal feature query through every matching string and procedure, their xrefs, and truthful containing-procedure resolution. Returns observations, operation count, and residual unknowns without inferring reference kinds.", enhancedInputSchemas.trace_feature, ), enhanced( @@ -392,12 +378,12 @@ export const SESSION_TOOL_CONTRACTS = [ ), session( "import_evidence_bundle", - "Read a bounded local JSON bundle beneath an operator-approved root, validate every Evidence v2 ID and canonical manifest, then atomically merge it. Imported content is data only and is never executed.", + "Read a bounded local JSON bundle beneath an operator-approved root, validate every Evidence ID and canonical manifest, then atomically merge it. Imported content is data only and is never executed.", importEvidenceBundleInputSchema, ), session( "capture_process_scenario", - "Run one bounded process under a PTY using operator-approved executable and working roots. Produces process capture Evidence; unknown_registry_approved: true separately records capture residuals. Captures raw and xterm-rendered terminal frames, scripted interactions, lifecycle filesystem checkpoints, process ownership, declarative command shims, and loopback replay. Disabled unless operator policy enables it; not a security sandbox.", + "Run one bounded process under a PTY using operator-approved executable and working roots. Produces process capture Evidence and records residual unknowns linked to that Evidence. Captures raw and xterm-rendered terminal frames, scripted interactions, lifecycle filesystem checkpoints, process ownership, declarative command shims, and loopback replay. Disabled unless operator policy enables it; not a security sandbox.", processScenarioSchema, ), session( diff --git a/src/contracts/toolEffects.ts b/src/contracts/toolEffects.ts index c4a6737dc2..40726e755f 100644 --- a/src/contracts/toolEffects.ts +++ b/src/contracts/toolEffects.ts @@ -80,7 +80,6 @@ export const TOOL_EFFECTS: Readonly> = { mayDiscardData: true, }), xrefs: evidence, - swift_classes: evidence, get_objc_classes: evidence, get_objc_protocols: evidence, batch_decompile: evidence, diff --git a/src/contracts/toolOutputSchemaGroups.ts b/src/contracts/toolOutputSchemaGroups.ts index 00975f8460..862764e4d6 100644 --- a/src/contracts/toolOutputSchemaGroups.ts +++ b/src/contracts/toolOutputSchemaGroups.ts @@ -207,7 +207,6 @@ const callPathTraceOutput = resultOf( /** Exact structured-content schemas for composed analysis workflows. */ export const enhancedOutputSchemas: Readonly> = { - swift_classes: symbolDiscoveryOutput("classes"), get_objc_classes: symbolDiscoveryOutput("classes"), get_objc_protocols: symbolDiscoveryOutput("protocols"), batch_decompile: resultOf( @@ -282,7 +281,7 @@ export const enhancedOutputSchemas: Readonly> = { trace_call_path: callPathTraceOutput, }; -/** Exact Evidence v2 schemas for provider-neutral native inspection. */ +/** Exact Evidence schemas for provider-neutral native inspection. */ export const nativeOutputSchemas: Readonly> = { inspect_macho: resultOf(inspectMachoSchema), inspect_signature: resultOf(inspectSignatureSchema), @@ -291,13 +290,13 @@ export const nativeOutputSchemas: Readonly> = { demangle_swift: resultOf(demangleSwiftSchema), }; -/** Exact Evidence v2 schemas for provider-neutral artifact graph operations. */ +/** Exact Evidence schemas for provider-neutral artifact graph operations. */ export const artifactOutputSchemas: Readonly> = { inspect_artifact: resultOf(artifactInspectionResultSchema), extract_artifact: resultOf(artifactExtractionResultSchema), }; -/** Exact Evidence v2 schema for execution-free managed static analysis. */ +/** Exact Evidence schema for execution-free managed static analysis. */ export const managedOutputSchemas: Readonly> = { inspect_managed_artifact: resultOf(managedArtifactInspectionSchema), inspect_managed_members: resultOf(managedMemberInspectionSchema), @@ -306,7 +305,7 @@ export const managedOutputSchemas: Readonly> = { ), }; -/** Exact Evidence v2 schema for provider-neutral managed workflows. */ +/** Exact Evidence schema for provider-neutral managed workflows. */ export const managedWorkflowOutputSchemas: Readonly< Record > = { diff --git a/src/contracts/toolSchemaMetadata.ts b/src/contracts/toolSchemaMetadata.ts index 14d43d805c..3f52fe37b6 100644 --- a/src/contracts/toolSchemaMetadata.ts +++ b/src/contracts/toolSchemaMetadata.ts @@ -17,8 +17,8 @@ const PROPERTY_DESCRIPTIONS: Readonly> = { direction: "Direction in which to traverse or compare relationships.", document: "Exact provider document or program identity.", error: "Structured, caller-actionable error when the operation fails.", - evidence: "Evidence v2 record produced by this operation.", - evidence_id: "Stable identifier of the recorded Evidence v2 observation.", + evidence: "Evidence record produced by this operation.", + evidence_id: "Stable identifier of the recorded Evidence observation.", executable: "Approved absolute path of the executable to run.", format: "Declared input artifact format.", name: "Exact name used by this operation.", @@ -44,8 +44,6 @@ const PROPERTY_DESCRIPTIONS: Readonly> = { symbols: "Ordered bounded Swift symbols to demangle.", target_id: "Exact authorized CDP target identifier.", unknown_id: "Exact residual-unknown identifier.", - unknown_registry_approved: - "Explicit approval to record bounded residual uncertainty in the session registry.", }; /** Attach caller guidance to a canonical schema for the SDK wire projection. */ @@ -124,9 +122,9 @@ const fallbackPropertyDescription = (property: string): string => { if (property.endsWith("_sha256")) return `Exact SHA-256 digest of ${words.slice(0, -7)}.`; if (property.endsWith("_evidence_id")) - return `Exact Evidence v2 identifier for the ${words.slice(0, -12)} observation.`; + return `Exact Evidence identifier for the ${words.slice(0, -12)} observation.`; if (property.endsWith("_evidence_ids")) - return `Ordered Evidence v2 identifiers for the ${words.slice(0, -13)} observations.`; + return `Ordered Evidence identifiers for the ${words.slice(0, -13)} observations.`; if (property.endsWith("_path")) return `Local filesystem path for ${words.slice(0, -5)}.`; if (property.endsWith("_uri")) diff --git a/src/domain/artifactComparison.ts b/src/domain/artifactComparison.ts index 35214354ba..297ee3961f 100644 --- a/src/domain/artifactComparison.ts +++ b/src/domain/artifactComparison.ts @@ -44,10 +44,6 @@ const comparisonDimensionSchema = z.enum([ export const artifactComparisonInputSchema = z.strictObject({ left_evidence_id: evidenceIdSchema, right_evidence_id: evidenceIdSchema, - unknown_registry_approved: z - .literal(true) - .optional() - .describe("Explicit approval to record incomplete or differing evidence"), }); /** One path-classified artifact change with citations to both observations. */ diff --git a/src/domain/browserScenario.ts b/src/domain/browserScenario.ts index 2546089f23..7d8a0fd524 100644 --- a/src/domain/browserScenario.ts +++ b/src/domain/browserScenario.ts @@ -2,10 +2,10 @@ import { z } from "zod"; import { BROWSER_SCENARIO_LIMITS, + BROWSER_SCENARIO_CAPTURE_LIMITS, browserScenarioActionSchema, browserScenarioAllowedOriginsSchema, browserScenarioBrowserSchema, - browserScenarioCaptureLimitsSchema, browserScenarioCaptureSchema, browserScenarioEnvironmentSchema, browserScenarioRedactionSchema, @@ -28,26 +28,42 @@ export { type BrowserScenarioValue, } from "./browserScenarioValues.js"; -const scenarioShapeSchema = z.strictObject({ - browser: browserScenarioBrowserSchema, - start_url: browserScenarioUrlSchema, - allowed_origins: browserScenarioAllowedOriginsSchema, +export const browserScenarioInputSchema = z.strictObject({ + browser: browserScenarioBrowserSchema.describe( + "Required launch/connect authority. Launch uses the approved executable and an owned temporary profile; connect uses one approved loopback CDP endpoint and target.", + ), + start_url: browserScenarioUrlSchema.describe( + "Required initial HTTP(S) URL. Query values must be declared separately as public literals or secret references.", + ), + allowed_origins: browserScenarioAllowedOriginsSchema.describe( + "Required exact HTTP(S) origins for the start URL and every later navigation, storage seed, and replay route.", + ), environment: browserScenarioEnvironmentSchema, actions: z .array(browserScenarioActionSchema) .min(1) - .max(BROWSER_SCENARIO_LIMITS.actions), + .max(BROWSER_SCENARIO_LIMITS.actions) + .describe("Required bounded sequence of explicit browser actions."), storage: browserScenarioStorageSchema, request_replay: browserScenarioRequestReplaySchema, secrets: z .array(browserScenarioSecretSchema) .max(BROWSER_SCENARIO_LIMITS.secrets) - .default([]), + .default([]) + .describe( + "Optional references to approved environment variables. Secret values are never supplied inline; declarations are required only when an action, URL, storage value, or replay route references a secret.", + ), redaction: browserScenarioRedactionSchema, capture: browserScenarioCaptureSchema, - limits: browserScenarioCaptureLimitsSchema, }); +const scenarioShapeSchema = browserScenarioInputSchema.transform( + (scenario) => ({ + ...scenario, + limits: BROWSER_SCENARIO_CAPTURE_LIMITS, + }), +); + type ScenarioShape = z.infer; const addIssue = ( diff --git a/src/domain/browserScenarioValues.ts b/src/domain/browserScenarioValues.ts index e53aeaa803..c7e5276280 100644 --- a/src/domain/browserScenarioValues.ts +++ b/src/domain/browserScenarioValues.ts @@ -15,6 +15,42 @@ export const BROWSER_SCENARIO_LIMITS = { actionTimeoutMs: 60_000, } as const; +/** Fixed internal budgets for provider-owned scenario capture. */ +export interface BrowserScenarioCaptureLimits { + readonly max_duration_ms: number; + readonly action_timeout_ms: number; + readonly navigation_timeout_ms: number; + readonly max_events: number; + readonly max_frames: number; + readonly max_workers: number; + readonly max_popups: number; + readonly max_websockets: number; + readonly max_dom_nodes: number; + readonly max_accessibility_nodes: number; + readonly max_screenshots: number; + readonly max_screenshot_bytes: number; + readonly max_storage_entries: number; + readonly max_total_metadata_bytes: number; +} + +export const BROWSER_SCENARIO_CAPTURE_LIMITS: BrowserScenarioCaptureLimits = + Object.freeze({ + max_duration_ms: 60_000, + action_timeout_ms: 5_000, + navigation_timeout_ms: 10_000, + max_events: 2_000, + max_frames: 100, + max_workers: 20, + max_popups: 10, + max_websockets: 100, + max_dom_nodes: 10_000, + max_accessibility_nodes: 10_000, + max_screenshots: 16, + max_screenshot_bytes: 4 * 1_024 * 1_024, + max_storage_entries: 256, + max_total_metadata_bytes: 4 * 1_024 * 1_024, + }); + export const scenarioIdentifierSchema = z .string() .regex(/^[A-Za-z][A-Za-z0-9._-]{0,63}$/u); @@ -102,26 +138,42 @@ export const browserScenarioBrowserSchema = z.discriminatedUnion("mode", [ }), ]); -export const browserScenarioEnvironmentSchema = z.strictObject({ - viewport: z.strictObject({ - width: z.number().int().min(320).max(7_680), - height: z.number().int().min(240).max(4_320), - device_scale_factor: z.number().min(0.5).max(4).default(1), - }), - locale: z - .string() - .min(2) - .max(64) - .regex(/^[A-Za-z0-9]+(?:-[A-Za-z0-9]+)*$/u), - timezone: z - .string() - .min(1) - .max(128) - .regex(/^[A-Za-z0-9_+-]+(?:\/[A-Za-z0-9_+-]+)*$/u), - color_scheme: z.enum(["light", "dark", "no-preference"]), - reduced_motion: z.enum(["reduce", "no-preference"]), - service_workers: z.literal("block"), -}); +export const browserScenarioEnvironmentSchema = z + .strictObject({ + viewport: z + .strictObject({ + width: z.number().int().min(320).max(7_680).default(1_280), + height: z.number().int().min(240).max(4_320).default(720), + device_scale_factor: z.number().min(0.5).max(4).default(1), + }) + .default({ width: 1_280, height: 720, device_scale_factor: 1 }), + locale: z + .string() + .min(2) + .max(64) + .regex(/^[A-Za-z0-9]+(?:-[A-Za-z0-9]+)*$/u) + .default("en-US"), + timezone: z + .string() + .min(1) + .max(128) + .regex(/^[A-Za-z0-9_+-]+(?:\/[A-Za-z0-9_+-]+)*$/u) + .default("UTC"), + color_scheme: z.enum(["light", "dark", "no-preference"]).default("light"), + reduced_motion: z.enum(["reduce", "no-preference"]).default("reduce"), + service_workers: z.literal("block").default("block"), + }) + .default({ + viewport: { width: 1_280, height: 720, device_scale_factor: 1 }, + locale: "en-US", + timezone: "UTC", + color_scheme: "light", + reduced_motion: "reduce", + service_workers: "block", + }) + .describe( + "Optional deterministic browser settings. Defaults to 1280x720, en-US, UTC, light mode, reduced motion, and blocked service workers.", + ); const locatorSchema = z.discriminatedUnion("kind", [ z.strictObject({ @@ -225,39 +277,44 @@ const storageEntrySchema = z.strictObject({ value: browserScenarioValueSchema, }); -export const browserScenarioStorageSchema = z.strictObject({ - cookies: z - .array( - z.strictObject({ - name: z.string().min(1).max(1_024), - value: browserScenarioValueSchema, - destination: browserScenarioUrlSchema, - http_only: z.boolean(), - secure: z.boolean(), - same_site: z.enum(["Strict", "Lax", "None"]), - }), - ) - .max(128) - .default([]), - local_storage: z - .array( - z.strictObject({ - origin: browserOriginSchema, - entries: z.array(storageEntrySchema).max(128), - }), - ) - .max(32) - .default([]), - session_storage: z - .array( - z.strictObject({ - origin: browserOriginSchema, - entries: z.array(storageEntrySchema).max(128), - }), - ) - .max(32) - .default([]), -}); +export const browserScenarioStorageSchema = z + .strictObject({ + cookies: z + .array( + z.strictObject({ + name: z.string().min(1).max(1_024), + value: browserScenarioValueSchema, + destination: browserScenarioUrlSchema, + http_only: z.boolean(), + secure: z.boolean(), + same_site: z.enum(["Strict", "Lax", "None"]), + }), + ) + .max(128) + .default([]), + local_storage: z + .array( + z.strictObject({ + origin: browserOriginSchema, + entries: z.array(storageEntrySchema).max(128), + }), + ) + .max(32) + .default([]), + session_storage: z + .array( + z.strictObject({ + origin: browserOriginSchema, + entries: z.array(storageEntrySchema).max(128), + }), + ) + .max(32) + .default([]), + }) + .default({ cookies: [], local_storage: [], session_storage: [] }) + .describe( + "Optional initial cookies, local storage, or session storage. Defaults to empty; every supplied value and origin must be explicitly declared and approved.", + ); const headerNameSchema = z .string() @@ -298,17 +355,22 @@ const replayRouteSchema = z.strictObject({ response: replayResponseSchema, }); -export const browserScenarioRequestReplaySchema = z.discriminatedUnion("mode", [ - z.strictObject({ mode: z.literal("disabled") }), - z.strictObject({ - mode: z.literal("exact"), - unmatched: z.enum(["abort", "passthrough-approved-origins"]), - routes: z - .array(replayRouteSchema) - .min(1) - .max(BROWSER_SCENARIO_LIMITS.replayRoutes), - }), -]); +export const browserScenarioRequestReplaySchema = z + .discriminatedUnion("mode", [ + z.strictObject({ mode: z.literal("disabled") }), + z.strictObject({ + mode: z.literal("exact"), + unmatched: z.enum(["abort", "passthrough-approved-origins"]), + routes: z + .array(replayRouteSchema) + .min(1) + .max(BROWSER_SCENARIO_LIMITS.replayRoutes), + }), + ]) + .default({ mode: "disabled" }) + .describe( + "Optional exact request replay. Defaults to disabled; replay routes and unmatched-request behavior must be explicitly declared when enabled.", + ); export const browserScenarioSecretSchema = z.strictObject({ secret_id: scenarioIdentifierSchema, @@ -339,7 +401,9 @@ const normalizedNames = (maximum: number) => export const browserScenarioRedactionSchema = z .strictObject({ - secret_values: z.literal("replace-with-secret-reference"), + secret_values: z + .literal("replace-with-secret-reference") + .default("replace-with-secret-reference"), query_parameter_names: normalizedNames(128).default([]), header_names: normalizedNames(128).default([...REQUIRED_REDACTED_HEADERS]), }) @@ -351,7 +415,15 @@ export const browserScenarioRedactionSchema = z path: ["header_names"], message: `Required credential header ${required} must be redacted`, }); - }); + }) + .default({ + secret_values: "replace-with-secret-reference", + query_parameter_names: [], + header_names: [...REQUIRED_REDACTED_HEADERS], + }) + .describe( + "Redaction policy. Secret values and credential headers are always redacted; list query parameter names when a secret is used in a URL.", + ); const snapshotKindSchema = z.enum([ "screenshot", @@ -362,62 +434,28 @@ const snapshotKindSchema = z.enum([ "storage", ]); -export const browserScenarioCaptureSchema = z.strictObject({ - after_each_step: z.array(snapshotKindSchema).max(6), - at_end: z.array(snapshotKindSchema).max(6), - events: z.array( - z.enum([ - "console", - "page-errors", - "network", - "websockets", - "frames", - "workers", - "popups", - "downloads", - ]), - ), -}); - -export const browserScenarioCaptureLimitsSchema = z.strictObject({ - max_duration_ms: z - .number() - .int() - .min(100) - .max(BROWSER_SCENARIO_LIMITS.durationMs), - action_timeout_ms: z - .number() - .int() - .min(1) - .max(BROWSER_SCENARIO_LIMITS.actionTimeoutMs), - navigation_timeout_ms: z - .number() - .int() - .min(1) - .max(BROWSER_SCENARIO_LIMITS.actionTimeoutMs), - max_events: z.number().int().min(1).max(100_000), - max_frames: z.number().int().min(1).max(1_000), - max_workers: z.number().int().min(0).max(1_000), - max_popups: z.number().int().min(0).max(100), - max_websockets: z.number().int().min(0).max(10_000), - max_dom_nodes: z.number().int().min(1).max(100_000), - max_accessibility_nodes: z.number().int().min(1).max(100_000), - max_screenshots: z.number().int().min(0).max(256), - max_screenshot_bytes: z - .number() - .int() - .min(1) - .max(8 * 1_024 * 1_024), - max_storage_entries: z - .number() - .int() - .min(0) - .max(BROWSER_SCENARIO_LIMITS.storageEntries), - max_total_metadata_bytes: z - .number() - .int() - .min(1_024) - .max(64 * 1_024 * 1_024), -}); +export const browserScenarioCaptureSchema = z + .strictObject({ + after_each_step: z.array(snapshotKindSchema).max(6).default([]), + at_end: z.array(snapshotKindSchema).max(6).default(["url"]), + events: z + .array( + z.enum([ + "console", + "page-errors", + "network", + "websockets", + "frames", + "workers", + "popups", + "downloads", + ]), + ) + .default([]), + }) + .default({ after_each_step: [], at_end: ["url"], events: [] }) + .describe( + "Optional retained artifacts and event families. Defaults to only a final sanitized URL; request screenshots, DOM, accessibility, history, storage, and event capture explicitly.", + ); export const browserScenarioAllowedOriginsSchema = browserAllowedOriginsSchema; diff --git a/src/domain/symbolAnalysis.ts b/src/domain/symbolAnalysis.ts index c1ec545830..a733ecc773 100644 --- a/src/domain/symbolAnalysis.ts +++ b/src/domain/symbolAnalysis.ts @@ -1,20 +1,6 @@ import type { JsonValue } from "./jsonValue.js"; import type { AddressedName } from "./hopperValues.js"; -/** Select Swift class symbols using the legacy `_TtC` convention. */ -export const discoverSwiftClasses = ( - procedures: readonly AddressedName[], - pattern: string, -): JsonValue => { - const classes = procedures - .filter(({ name }) => name.includes("_TtC")) - .filter(({ name }) => pattern.length === 0 || name.includes(pattern)); - return { - count: classes.length, - classes: classes.map(toJsonEntry), - }; -}; - /** Select and deduplicate Objective-C class labels. */ export const discoverObjcClasses = ( names: readonly AddressedName[], @@ -55,10 +41,15 @@ const SWIFT_CATEGORIES = [ ["protocols", "_TtP"], ["extensions", "_TtE"], ] as const; +type SwiftTypeCategory = (typeof SWIFT_CATEGORIES)[number][0] | "other"; /** Categorize deduplicated Swift mangled symbols. */ export const categorizeSwiftTypes = ( procedures: readonly AddressedName[], + filter: { + readonly category?: SwiftTypeCategory | undefined; + readonly pattern?: string | undefined; + } = {}, ): JsonValue => { const groups: Record = Object.fromEntries( [...SWIFT_CATEGORIES.map(([category]) => category), "other"].map( @@ -69,11 +60,14 @@ export const categorizeSwiftTypes = ( for (const entry of procedures) { if (!entry.name.includes("_Tt") || seen.has(entry.name)) continue; - seen.add(entry.name); const category = SWIFT_CATEGORIES.find(([, prefix]) => entry.name.startsWith(prefix), )?.[0] ?? "other"; + if (filter.category !== undefined && filter.category !== category) continue; + if (filter.pattern !== undefined && !entry.name.includes(filter.pattern)) + continue; + seen.add(entry.name); groups[category]?.push(entry); } diff --git a/src/domain/webContentArtifact.ts b/src/domain/webContentArtifact.ts index 308d4d8749..6a063ad9f2 100644 --- a/src/domain/webContentArtifact.ts +++ b/src/domain/webContentArtifact.ts @@ -5,7 +5,6 @@ import { z } from "zod"; /** Self-verifying text artifact used by web-source analysis boundaries. */ export const webTextArtifactSchema = z .object({ - uri: z.string().regex(/^rea:\/\/web-content\/sha256\/[a-f0-9]{64}$/u), sha256: z.string().regex(/^[a-f0-9]{64}$/u), bytes: z.number().int().min(0), media_type: z.string().min(1).max(256), @@ -27,12 +26,6 @@ export const webTextArtifactSchema = z path: ["sha256"], message: "Digest mismatch", }); - if (artifact.uri !== webContentUri(digest)) - context.addIssue({ - code: "custom", - path: ["uri"], - message: "Content URI mismatch", - }); }); export type WebTextArtifact = z.infer; @@ -43,7 +36,6 @@ export const createWebTextArtifact = ( ): WebTextArtifact => { const sha256 = createHash("sha256").update(text).digest("hex"); return { - uri: webContentUri(sha256), sha256, bytes: Buffer.byteLength(text), media_type: mediaType, @@ -51,6 +43,3 @@ export const createWebTextArtifact = ( text, }; }; - -const webContentUri = (sha256: string): string => - `rea://web-content/sha256/${sha256}`; diff --git a/src/domain/webScreenshot.ts b/src/domain/webScreenshot.ts index 587eddc298..eb26311719 100644 --- a/src/domain/webScreenshot.ts +++ b/src/domain/webScreenshot.ts @@ -13,7 +13,6 @@ const MAX_SCREENSHOT_BYTES = 8 * 1_024 * 1_024; /** Self-verifying inline PNG artifact for CLI/MCP parity. */ export const webScreenshotArtifactSchema = z .object({ - uri: z.string().regex(/^rea:\/\/web-screenshot\/sha256\/[a-f0-9]{64}$/u), sha256: z.string().regex(/^[a-f0-9]{64}$/u), bytes: z.number().int().min(1).max(MAX_SCREENSHOT_BYTES), media_type: z.literal("image/png"), @@ -26,11 +25,7 @@ export const webScreenshotArtifactSchema = z return; } const sha256 = createHash("sha256").update(bytes).digest("hex"); - if ( - bytes.byteLength !== artifact.bytes || - sha256 !== artifact.sha256 || - artifact.uri !== `rea://web-screenshot/sha256/${sha256}` - ) + if (bytes.byteLength !== artifact.bytes || sha256 !== artifact.sha256) context.addIssue({ code: "custom", message: "Screenshot artifact digest or size mismatch", @@ -164,7 +159,6 @@ export const createWebScreenshotArtifact = ( ): WebScreenshotArtifact => { const sha256 = createHash("sha256").update(bytes).digest("hex"); return webScreenshotArtifactSchema.parse({ - uri: `rea://web-screenshot/sha256/${sha256}`, sha256, bytes: bytes.byteLength, media_type: "image/png", diff --git a/src/server/ProcessCaptureElicitation.ts b/src/server/ProcessCaptureElicitation.ts index 7d1e50b8cc..2d8dcc6c29 100644 --- a/src/server/ProcessCaptureElicitation.ts +++ b/src/server/ProcessCaptureElicitation.ts @@ -19,7 +19,6 @@ import { } from "../domain/permissionPolicy.js"; const RESPONSE_KEY = "process_capture_grant"; -const STATE_VERSION = 1; /** Shared continuation, replay, and round limits for process capture consent. */ export const PROCESS_CAPTURE_ELICITATION_POLICY = { @@ -41,7 +40,6 @@ Reflect.deleteProperty(grantResponseJsonSchema, "$schema"); Reflect.deleteProperty(grantResponseJsonSchema, "additionalProperties"); const stateSchema = z.strictObject({ - version: z.literal(STATE_VERSION), nonce: z.string().uuid(), request: z.strictObject({ capability: z.literal("process_capture"), @@ -130,7 +128,6 @@ export const authorizeProcessCaptureWithElicitation = async ( } const state = await elicitation.stateCodec.mint({ - version: STATE_VERSION, nonce: randomUUID(), request: { capability: "process_capture", diff --git a/src/server/createServer.ts b/src/server/createServer.ts index cab2eaded1..fd706c5b99 100644 --- a/src/server/createServer.ts +++ b/src/server/createServer.ts @@ -58,7 +58,7 @@ const TARGET_FREE_INSTRUCTIONS = "Start with binary_session to see the active target and available providers. For an archive or application package, call open_binary(path), then inspect_artifact for the complete graph and findings; use extract_artifact only for selected files. For standalone JavaScript or ASAR analysis, use analyze_javascript_application. For managed PE/CLI files, use inspect_managed_artifact. For native binaries, open the target and start with binary_overview. For a live page or runtime, discover targets with list_browser_targets, list_electron_targets, or list_javascript_runtime_targets before inspection; scenario capture tools are for controlled interactions."; const ACTIVE_TARGET_INSTRUCTIONS = - "REA analyzes the active reverse-engineering target. Start native analysis with binary_overview, then use analyze_function, literal search, callers, callees, and xrefs as needed."; + "REA analyzes the active reverse-engineering target. For native binaries, call binary_overview for target metadata and inventory counts, then search_procedures or list_procedures to choose a function and analyze_function with its returned name or address for a complete dossier. Use inspect_native_api when you need to reconstruct a native API boundary, and focused procedure tools when you need only one facet such as assembly, callers, or references."; export interface CreateServerOptions { readonly logger?: Logger; diff --git a/src/server/recordDerivedEvidence.ts b/src/server/recordDerivedEvidence.ts index 99a761c425..877be2acf3 100644 --- a/src/server/recordDerivedEvidence.ts +++ b/src/server/recordDerivedEvidence.ts @@ -4,7 +4,7 @@ import type { Evidence } from "../domain/evidence.js"; import type { RecordUnknownInput } from "../domain/residualUnknown.js"; import { ok, type Result } from "../domain/result.js"; -/** Atomically record derived Evidence with an optional approved unknown. */ +/** Atomically record derived Evidence with an optional supported unknown. */ export const recordDerivedEvidence = ( session: BinarySessionPort, evidence: Evidence, diff --git a/src/server/registerApplicationTools/compareExportShapes.ts b/src/server/registerApplicationTools/compareExportShapes.ts index b2c2397db0..29a8beaea1 100644 --- a/src/server/registerApplicationTools/compareExportShapes.ts +++ b/src/server/registerApplicationTools/compareExportShapes.ts @@ -45,9 +45,7 @@ export const registerCompareJavaScriptExportShapesTool = ( options, contract, result.value, - parsed.unknown_registry_approved === true && unknown - ? "javascript-export-shape" - : undefined, + unknown ? "javascript-export-shape" : undefined, ); }, ); diff --git a/src/server/registerApplicationTools/compareSourceToBundle.ts b/src/server/registerApplicationTools/compareSourceToBundle.ts index a9948bb81a..8ccaf4b85f 100644 --- a/src/server/registerApplicationTools/compareSourceToBundle.ts +++ b/src/server/registerApplicationTools/compareSourceToBundle.ts @@ -42,8 +42,7 @@ export const registerCompareSourceToBundleTool = ( options, contract, result.value, - resolved.value.unknown_registry_approved === true && - comparison.summary.unknown > 0 + comparison.summary.unknown > 0 ? "source-to-bundle-comparison" : undefined, ); diff --git a/src/server/registerApplicationTools/compareVersions.ts b/src/server/registerApplicationTools/compareVersions.ts index 661e8adbad..f79700289c 100644 --- a/src/server/registerApplicationTools/compareVersions.ts +++ b/src/server/registerApplicationTools/compareVersions.ts @@ -54,9 +54,7 @@ export const registerCompareApplicationVersionsTool = ( options, compareContract, result.value, - parsed.unknown_registry_approved === true && unknown - ? "application-version-comparison" - : undefined, + unknown ? "application-version-comparison" : undefined, ); }, ); diff --git a/src/server/registerArtifactComparisonTool.ts b/src/server/registerArtifactComparisonTool.ts index 1294a01f41..b1ae6451d8 100644 --- a/src/server/registerArtifactComparisonTool.ts +++ b/src/server/registerArtifactComparisonTool.ts @@ -75,11 +75,13 @@ const artifactUnknownInput = ( input: { readonly left_evidence_id: string; readonly right_evidence_id: string; - readonly unknown_registry_approved?: true | undefined; }, status: ReturnType["status"], ): RecordUnknownInput | undefined => { - if (input.unknown_registry_approved !== true || status === "unchanged") + if ( + status === "unchanged" || + input.left_evidence_id === input.right_evidence_id + ) return undefined; return { question: `Artifact comparison is ${status}`, diff --git a/src/server/registerBrowserScenarioTool.ts b/src/server/registerBrowserScenarioTool.ts index a60d712469..86464a5d6c 100644 --- a/src/server/registerBrowserScenarioTool.ts +++ b/src/server/registerBrowserScenarioTool.ts @@ -5,6 +5,7 @@ import type { BrowserScenarioCapturePort } from "../application/BrowserScenarioC import { captureBrowserScenario } from "../application/BrowserScenarioCaptureService.js"; import type { PermissionAuthority } from "../application/PermissionAuthority.js"; import { BROWSER_SCENARIO_TOOL_CONTRACTS } from "../contracts/browserScenarioToolContracts.js"; +import { browserScenarioSchema } from "../domain/browserScenario.js"; import type { Logger } from "../logger.js"; import { logToolExecution } from "./toolLogging.js"; import { toolRegistrationOptions } from "./toolRegistrationOptions.js"; @@ -27,11 +28,12 @@ export const registerBrowserScenarioTool = ( contract.name, toolRegistrationOptions(contract), async (input, context) => { + const scenario = browserScenarioSchema.parse(input); const result = await logToolExecution(options.logger, contract.name, () => captureBrowserScenario( options.provider, options.permissionAuthority, - input, + scenario, { signal: context.mcpReq.signal }, ), ); diff --git a/src/server/registerEnhancedTools.ts b/src/server/registerEnhancedTools.ts index 27d3dece0d..bb3fab38c0 100644 --- a/src/server/registerEnhancedTools.ts +++ b/src/server/registerEnhancedTools.ts @@ -48,7 +48,6 @@ export const registerEnhancedTools = ( options: EnhancedToolRegistration, ): void => { const [ - swiftClasses, objcClasses, objcProtocols, batchDecompile, @@ -62,15 +61,6 @@ export const registerEnhancedTools = ( codeForString, traceCallPath, ] = ENHANCED_TOOL_CONTRACTS; - server.registerTool( - swiftClasses.name, - toolRegistrationOptions(swiftClasses), - (input, context) => - executeEnhancedTool(analysis, options, swiftClasses, { - validatedCall: { name: "swift_classes", input }, - context, - }), - ); server.registerTool( objcClasses.name, toolRegistrationOptions(objcClasses), @@ -243,7 +233,6 @@ const executeEnhancedTool = async ( return toCallToolResult(recorded, contract); const unknowns = recordWorkflowUnknowns({ name, - input: parameters, result: result.value, evidenceId: evidence.evidence_id, recordUnknown: registration.recordUnknown, @@ -265,7 +254,6 @@ const jsonParameters = ( interface WorkflowUnknownInput { readonly name: string; - readonly input: Readonly>; readonly result: JsonValue; readonly evidenceId: string; readonly recordUnknown: BinarySessionPort["recordUnknown"] | undefined; @@ -273,7 +261,6 @@ interface WorkflowUnknownInput { const recordWorkflowUnknowns = ({ name, - input, result, evidenceId, recordUnknown, @@ -287,7 +274,6 @@ const recordWorkflowUnknowns = ({ "trace_call_path", "inspect_native_api", ].includes(name) || - input.unknown_registry_approved !== true || recordUnknown === undefined || typeof result !== "object" || result === null || diff --git a/src/server/registerFunctionComparisonTool.ts b/src/server/registerFunctionComparisonTool.ts index f95756d2bd..9bc54ba252 100644 --- a/src/server/registerFunctionComparisonTool.ts +++ b/src/server/registerFunctionComparisonTool.ts @@ -67,7 +67,6 @@ export const registerFunctionComparisonTool = ( session, evidence, functionUnknownInput({ - approved: input.unknown_registry_approved, status: comparison.status, leftIds, rightIds, @@ -79,17 +78,15 @@ export const registerFunctionComparisonTool = ( }; const functionUnknownInput = ({ - approved, status, leftIds, rightIds, }: { - approved: true | undefined; status: ReturnType["status"]; leftIds: readonly string[]; rightIds: readonly string[]; }): RecordUnknownInput | undefined => { - if (approved !== true || status === "unchanged") return undefined; + if (status === "unchanged") return undefined; return { question: `Function comparison is ${status}`, severity: status === "changed" ? "medium" : "high", diff --git a/src/server/registerInvestigationTools.ts b/src/server/registerInvestigationTools.ts index c3385a4246..d9feb1714a 100644 --- a/src/server/registerInvestigationTools.ts +++ b/src/server/registerInvestigationTools.ts @@ -79,7 +79,6 @@ const registerChangedBehavior = ( const recorded = recordWorkflowEvidence( session, evidence, - input.unknown_registry_approved, isIncomplete(result.behavior_status), { question: @@ -137,7 +136,6 @@ const registerCallPath = ( const recorded = recordWorkflowEvidence( session, evidence, - input.unknown_registry_approved, result.status === "unknown", { question: @@ -198,7 +196,6 @@ const registerStaticRuntime = ( recordWorkflowEvidence( session, evidence, - input.unknown_registry_approved, result.status === "unknown" || result.status === "truncated", { question: @@ -253,24 +250,16 @@ const registerReconstruction = ( evidenceLinks: links, }); return toCallToolResult( - recordWorkflowEvidence( - session, - evidence, - input.unknown_registry_approved, - result.status === "unknown", - { - question: "Does the reconstruction satisfy every declared claim?", - domain: "reconstruction-verification", - requiredAuthority: null, - requiredConfidence: "derived", - probes: result.recommended_probes.map( - ({ operation, rationale }) => ({ - operation, - rationale, - }), - ), - }, - ), + recordWorkflowEvidence(session, evidence, result.status === "unknown", { + question: "Does the reconstruction satisfy every declared claim?", + domain: "reconstruction-verification", + requiredAuthority: null, + requiredConfidence: "derived", + probes: result.recommended_probes.map(({ operation, rationale }) => ({ + operation, + rationale, + })), + }), contract, ); }, diff --git a/src/server/registerInvestigationTools/helpers.ts b/src/server/registerInvestigationTools/helpers.ts index c30ebcb100..602b1e5b7b 100644 --- a/src/server/registerInvestigationTools/helpers.ts +++ b/src/server/registerInvestigationTools/helpers.ts @@ -53,15 +53,14 @@ export const evidenceClosure = ( }; export const recordWorkflowEvidence = ( - ...[session, evidence, approved, unresolved, input]: readonly [ + ...[session, evidence, unresolved, input]: readonly [ session: BinarySessionPort, evidence: Evidence, - approved: true | undefined, unresolved: boolean, input: WorkflowUnknownInput, ] ): Result => { - if (approved !== true || !unresolved) { + if (!unresolved) { return recordDerivedEvidence(session, evidence, undefined); } return recordDerivedEvidence(session, evidence, { diff --git a/src/server/registerInvestigationTools/types.ts b/src/server/registerInvestigationTools/types.ts index e2fd1bbe64..4a652f15b4 100644 --- a/src/server/registerInvestigationTools/types.ts +++ b/src/server/registerInvestigationTools/types.ts @@ -1,4 +1,4 @@ -/** Fields used to build an approved residual unknown from a workflow result. */ +/** Fields used to build a residual unknown from a workflow result. */ export interface WorkflowUnknownInput { readonly question: string; readonly domain: string; diff --git a/src/server/registerManagedWorkflowTools/compareManagedMembers.ts b/src/server/registerManagedWorkflowTools/compareManagedMembers.ts index d4199743d3..c5b7f5b83b 100644 --- a/src/server/registerManagedWorkflowTools/compareManagedMembers.ts +++ b/src/server/registerManagedWorkflowTools/compareManagedMembers.ts @@ -50,28 +50,27 @@ export const registerCompareManagedMembers = ( result.value.normalized_result, ); const unknown = comparison.summary.unknown > 0; - const output = - parsed.unknown_registry_approved === true && unknown - ? options.recordEvidenceWithUnknown?.(result.value, { - question: - "Which managed members remain unmatched or ambiguous across these versions?", - severity: "medium", - domain: "managed-member-comparison", - supporting_evidence_ids: [result.value.evidence_id], - contradicting_evidence_ids: [], - required_authority: "shipped-artifact", - required_confidence: "observed", - required_environment: null, - recommended_probes: [ - { - operation: "inspect_managed_members", - rationale: - "Repeat static member inspection with complete pages and method bodies.", - }, - ], - relationships: [], - }) - : options.recordEvidence?.(result.value); + const output = unknown + ? options.recordEvidenceWithUnknown?.(result.value, { + question: + "Which managed members remain unmatched or ambiguous across these versions?", + severity: "medium", + domain: "managed-member-comparison", + supporting_evidence_ids: [result.value.evidence_id], + contradicting_evidence_ids: [], + required_authority: "shipped-artifact", + required_confidence: "observed", + required_environment: null, + recommended_probes: [ + { + operation: "inspect_managed_members", + rationale: + "Repeat static member inspection with complete pages and method bodies.", + }, + ], + relationships: [], + }) + : options.recordEvidence?.(result.value); if (output !== undefined && !output.ok) return toCallToolResult(output, compareContract); return toCallToolResult( diff --git a/src/server/registerManagedWorkflowTools/verifyManagedNativeBoundaries.ts b/src/server/registerManagedWorkflowTools/verifyManagedNativeBoundaries.ts index b0b5a31887..c6c8217727 100644 --- a/src/server/registerManagedWorkflowTools/verifyManagedNativeBoundaries.ts +++ b/src/server/registerManagedWorkflowTools/verifyManagedNativeBoundaries.ts @@ -71,33 +71,32 @@ export const registerVerifyManagedNativeBoundaries = ( verification.summary.unresolved > 0 || verification.summary.contradicted > 0 || verification.summary.native_body_unresolved > 0; - const output = - parsed.unknown_registry_approved === true && unknown - ? options.recordEvidenceWithUnknown?.(result.value, { - question: - "Which managed/native boundaries remain unresolved or contradicted by the supplied native Evidence?", - severity: "medium", - domain: "managed-native-verification", - supporting_evidence_ids: [result.value.evidence_id], - contradicting_evidence_ids: [], - required_authority: "shipped-artifact", - required_confidence: "observed", - required_environment: null, - recommended_probes: [ - { - operation: "inspect_managed_native_boundaries", - rationale: - "Repeat managed boundary inspection with complete ModuleRef, ImplMap, and native implementation pages.", - }, - { - operation: "analyze_function", - rationale: - "Analyze the provider-resolved native function candidate for the declared export.", - }, - ], - relationships: [], - }) - : options.recordEvidence?.(result.value); + const output = unknown + ? options.recordEvidenceWithUnknown?.(result.value, { + question: + "Which managed/native boundaries remain unresolved or contradicted by the supplied native Evidence?", + severity: "medium", + domain: "managed-native-verification", + supporting_evidence_ids: [result.value.evidence_id], + contradicting_evidence_ids: [], + required_authority: "shipped-artifact", + required_confidence: "observed", + required_environment: null, + recommended_probes: [ + { + operation: "inspect_managed_native_boundaries", + rationale: + "Repeat managed boundary inspection with complete ModuleRef, ImplMap, and native implementation pages.", + }, + { + operation: "analyze_function", + rationale: + "Analyze the provider-resolved native function candidate for the declared export.", + }, + ], + relationships: [], + }) + : options.recordEvidence?.(result.value); if (output !== undefined && !output.ok) return toCallToolResult(output, nativeVerificationContract); return toCallToolResult(result, nativeVerificationContract); diff --git a/src/server/registerOfficialTools.ts b/src/server/registerOfficialTools.ts index cc66f0bc5f..5ba620ea0e 100644 --- a/src/server/registerOfficialTools.ts +++ b/src/server/registerOfficialTools.ts @@ -1,4 +1,4 @@ -import type { McpServer } from "@modelcontextprotocol/server"; +import type { McpServer, ServerContext } from "@modelcontextprotocol/server"; import { z } from "zod"; import type { @@ -10,11 +10,7 @@ import type { ProgressReporter } from "../application/ProgressReporter.js"; import type { ToolContract } from "../contracts/toolContracts.js"; import { OFFICIAL_TOOL_CONTRACTS } from "../contracts/toolContracts.js"; import type { BinaryTarget } from "../domain/binaryTarget.js"; -import { - AnalysisCapabilityUnavailableError, - UnknownRegistryError, - type AnalysisError, -} from "../domain/errors.js"; +import type { AnalysisError } from "../domain/errors.js"; import { createEvidence } from "../domain/evidence.js"; import { jsonObjectSchema, @@ -33,7 +29,6 @@ export interface OfficialToolRegistration { readonly logger: Logger; readonly activeTarget: (() => BinaryTarget | undefined) | undefined; readonly recordEvidence: BinarySessionPort["recordEvidence"] | undefined; - readonly recordUnknown: BinarySessionPort["recordUnknown"] | undefined; } /** Register direct bridge proxies, preserving MCP cancellation and typed errors. */ @@ -47,7 +42,6 @@ export const registerOfficialTools = ( logger: options.logger, activeTarget: options.activeTarget, recordEvidence: options.recordEvidence, - recordUnknown: options.recordUnknown, }); } }; @@ -60,13 +54,12 @@ const registerOfficialTool = ( readonly logger: Logger; readonly activeTarget: (() => BinaryTarget | undefined) | undefined; readonly recordEvidence: BinarySessionPort["recordEvidence"] | undefined; - readonly recordUnknown: BinarySessionPort["recordUnknown"] | undefined; }, ): void => { server.registerTool( contract.name, toolRegistrationOptions(contract), - async (input, context) => { + async (input: unknown, context: ServerContext) => { const arguments_ = projectOfficialArguments(contract, input); const progress = mcpProgressReporter(context); const result = await runOfficialOperation( @@ -80,14 +73,6 @@ const registerOfficialTool = ( }, ); if (!result.ok) { - const unknownError = recordOfficialUnknown( - input, - result.error, - contract, - registration.recordUnknown, - ); - if (unknownError !== undefined) - return toCallToolResult(unknownError, contract); return toCallToolResult(result, contract); } const evidence = createEvidence( @@ -145,47 +130,6 @@ const runOfficialOperation = async ( return result; }; -const recordOfficialUnknown = ( - input: unknown, - error: AnalysisError, - contract: ToolContract, - recordUnknown: BinarySessionPort["recordUnknown"] | undefined, -): Result | undefined => { - if ( - !(error instanceof AnalysisCapabilityUnavailableError) || - !approvedUnknownTracking(input) || - recordUnknown === undefined - ) - return undefined; - const unknown = recordUnknown({ - question: "The requested analysis is unavailable for the current target.", - severity: "medium", - domain: "analysis-capability", - supporting_evidence_ids: [], - contradicting_evidence_ids: [], - required_authority: "shipped-artifact", - required_confidence: "observed", - required_environment: null, - recommended_probes: [ - { - operation: contract.name, - rationale: - "Choose another analysis or target that can answer this question.", - }, - ], - relationships: [], - }); - if ( - !unknown.ok && - !( - unknown.error instanceof UnknownRegistryError && - unknown.error.reason === "already-exists" - ) - ) - return unknown; - return undefined; -}; - const projectOfficialArguments = ( contract: ToolContract, input: unknown, @@ -198,14 +142,7 @@ const projectOfficialArguments = ( const projected: Record = {}; for (const key of Object.keys(contract.inputSchema.shape)) { - if (key === "unknown_registry_approved") continue; projected[key] = jsonValueSchema.parse(parsed[key] ?? null); } return projected; }; - -const approvedUnknownTracking = (input: unknown): boolean => - typeof input === "object" && - input !== null && - "unknown_registry_approved" in input && - input.unknown_registry_approved === true; diff --git a/src/server/registerProcessComparisonTool.ts b/src/server/registerProcessComparisonTool.ts index 8d36955748..215a268329 100644 --- a/src/server/registerProcessComparisonTool.ts +++ b/src/server/registerProcessComparisonTool.ts @@ -124,15 +124,10 @@ const comparisonUnknownInput = ( parsed: { readonly left_evidence_id: string; readonly right_evidence_id: string; - readonly unknown_registry_approved?: true | undefined; }, comparison: ReturnType, ): RecordUnknownInput | undefined => { - if ( - parsed.unknown_registry_approved !== true || - comparison.status === "unchanged" - ) - return undefined; + if (comparison.status === "unchanged") return undefined; const differingScopes = [ ["terminal", comparison.terminal], ["interaction", comparison.interaction], diff --git a/src/server/registerSessionTools.ts b/src/server/registerSessionTools.ts index 6ead11e93c..e2bd73c39b 100644 --- a/src/server/registerSessionTools.ts +++ b/src/server/registerSessionTools.ts @@ -60,7 +60,6 @@ const recordProcessResidualUnknowns = ( evidence: Evidence, residuals: ProcessCapture["residual_unknowns"], ): Result => { - if (scenario.unknown_registry_approved !== true) return ok(null); for (const residual of residuals) { const unknown = session.recordUnknown({ question: `Was ${residual.scope} behavior fully observed during capture?`, diff --git a/tests/acceptance/applications/applicationEvidenceMcp.test.ts b/tests/acceptance/applications/applicationEvidenceMcp.test.ts index 09d7c1537e..6faaa789ec 100644 --- a/tests/acceptance/applications/applicationEvidenceMcp.test.ts +++ b/tests/acceptance/applications/applicationEvidenceMcp.test.ts @@ -65,7 +65,6 @@ async function runInlineEvidenceScenarios( name: "compare_application_versions", arguments: { ...JAVASCRIPT_VERSION_COMPARISON_FULL_EVIDENCE_EXAMPLE, - unknown_registry_approved: true, }, }); expect(compared.isError).not.toBe(true); diff --git a/tests/acceptance/applications/applicationExportShapeMcp.test.ts b/tests/acceptance/applications/applicationExportShapeMcp.test.ts index 99e76a45c4..d6ec76e3ff 100644 --- a/tests/acceptance/applications/applicationExportShapeMcp.test.ts +++ b/tests/acceptance/applications/applicationExportShapeMcp.test.ts @@ -118,7 +118,7 @@ describe("application workflow MCP parity", () => { result: { source_evidence_id: left.value.evidence_id, source_graph_id: analyzed.semantic_graph.graph_id, - summary: { retained_seed_matches: 1 }, + summary: { total_seed_matches: 1 }, }, }); } finally { diff --git a/tests/boundary/browser/cdpBrowserProvider.document_script.test.ts b/tests/boundary/browser/cdpBrowserProvider.document_script.test.ts index 68d3f9a2b7..08d843b404 100644 --- a/tests/boundary/browser/cdpBrowserProvider.document_script.test.ts +++ b/tests/boundary/browser/cdpBrowserProvider.document_script.test.ts @@ -61,7 +61,7 @@ describeBrowser("CdpBrowserProvider: document script 1", () => { artifact: expect.objectContaining({ text: "export const observed = 'source-secret';", bytes: 40, - uri: expect.stringMatching(/^rea:\/\/web-content\/sha256\//u), + sha256: expect.stringMatching(/^[a-f0-9]{64}$/u), }), }), ); @@ -297,7 +297,7 @@ describeBrowser("CdpBrowserProvider: document script 3", () => { expect(result.value).toMatchObject({ viewport: { width: 1, height: 1 }, artifact: { - uri: expect.stringMatching(/^rea:\/\/web-screenshot\/sha256\//u), + sha256: expect.stringMatching(/^[a-f0-9]{64}$/u), bytes: 70, media_type: "image/png", }, diff --git a/tests/boundary/cli/browserCli.test.ts b/tests/boundary/cli/browserCli.test.ts index b3f047230c..7a92bbebb0 100644 --- a/tests/boundary/cli/browserCli.test.ts +++ b/tests/boundary/cli/browserCli.test.ts @@ -18,7 +18,7 @@ afterEach(async () => { describe("browser CLI parity", () => { it( - "returns the same Evidence v2 discovery and inspection contracts", + "returns the same Evidence discovery and inspection contracts", async () => { const browser = await startFakeCdpBrowser({ sessionTimeline: "same_origin", diff --git a/tests/boundary/mcp/artifactMcp.test.ts b/tests/boundary/mcp/artifactMcp.test.ts index cc2c3d0510..d84e5755cf 100644 --- a/tests/boundary/mcp/artifactMcp.test.ts +++ b/tests/boundary/mcp/artifactMcp.test.ts @@ -407,7 +407,6 @@ it("returns full artifact graphs inline and compares changed inventories", async arguments: { left_evidence_id: evidence.evidence_id, right_evidence_id: changedEvidence.evidence_id, - unknown_registry_approved: true, }, }); expect(compared.isError).not.toBe(true); diff --git a/tests/boundary/mcp/browserMcp.test.ts b/tests/boundary/mcp/browserMcp.test.ts index 0aa41555d2..dc4a218770 100644 --- a/tests/boundary/mcp/browserMcp.test.ts +++ b/tests/boundary/mcp/browserMcp.test.ts @@ -27,7 +27,7 @@ afterEach(async () => { }); it( - "exposes CLI-equivalent Evidence v2 results and session tools", + "exposes CLI-equivalent Evidence results and session tools", async () => { const browser = await startFakeCdpBrowser({ sessionTimeline: "same_origin", diff --git a/tests/boundary/mcp/browserScenarioMcp.test.ts b/tests/boundary/mcp/browserScenarioMcp.test.ts index 3ce2ffa648..35a7ddda3a 100644 --- a/tests/boundary/mcp/browserScenarioMcp.test.ts +++ b/tests/boundary/mcp/browserScenarioMcp.test.ts @@ -106,7 +106,7 @@ const captureFor = (scenario: BrowserScenario): BrowserScenarioCapture => { }); }; -const scenario = (origin = "https://app.example.test") => ({ +const minimalScenario = (origin = "https://app.example.test") => ({ browser: { mode: "launch", executable_path: process.execPath, @@ -116,14 +116,17 @@ const scenario = (origin = "https://app.example.test") => ({ }, start_url: { url: `${origin}/`, query: [] }, allowed_origins: [origin], - environment: { - viewport: { width: 1_280, height: 720 }, - locale: "en-US", - timezone: "UTC", - color_scheme: "light", - reduced_motion: "reduce", - service_workers: "block", - }, + actions: [ + { + step_id: "settle", + action: "wait_for_timeout", + duration_ms: 1, + }, + ], +}); + +const scenario = (origin = "https://app.example.test") => ({ + ...minimalScenario(origin), actions: [ { step_id: "login-input", @@ -132,8 +135,6 @@ const scenario = (origin = "https://app.example.test") => ({ value: { source: "secret", secret_id: "login_input" }, }, ], - storage: {}, - request_replay: { mode: "disabled" }, secrets: [ { secret_id: "login_input", @@ -142,27 +143,6 @@ const scenario = (origin = "https://app.example.test") => ({ redaction: "replace-with-secret-reference", }, ], - redaction: { - secret_values: "replace-with-secret-reference", - query_parameter_names: [], - }, - capture: { after_each_step: [], at_end: [], events: [] }, - limits: { - max_duration_ms: 10_000, - action_timeout_ms: 1_000, - navigation_timeout_ms: 1_000, - max_events: 100, - max_frames: 10, - max_workers: 5, - max_popups: 2, - max_websockets: 10, - max_dom_nodes: 1_000, - max_accessibility_nodes: 1_000, - max_screenshots: 2, - max_screenshot_bytes: 1_048_576, - max_storage_entries: 100, - max_total_metadata_bytes: 1_048_576, - }, }); describe("browser scenario MCP tool", () => { @@ -209,6 +189,28 @@ describe("browser scenario MCP tool", () => { await server.connect(serverTransport); await client.connect(clientTransport); + const minimal = await client.callTool({ + name: "capture_browser_scenario", + arguments: minimalScenario(), + }); + expect(minimal.isError, JSON.stringify(minimal)).not.toBe(true); + expect(provider.scenarios[0]).toMatchObject({ + environment: { color_scheme: "light", service_workers: "block" }, + storage: { cookies: [], local_storage: [], session_storage: [] }, + request_replay: { mode: "disabled" }, + redaction: { + secret_values: "replace-with-secret-reference", + header_names: [ + "authorization", + "cookie", + "proxy-authorization", + "set-cookie", + ], + }, + capture: { after_each_step: [], at_end: ["url"], events: [] }, + limits: { max_duration_ms: 60_000, max_total_metadata_bytes: 4_194_304 }, + }); + const captured = await client.callTool({ name: "capture_browser_scenario", arguments: scenario(), @@ -221,7 +223,7 @@ describe("browser scenario MCP tool", () => { }, }, }); - expect(provider.scenarios).toHaveLength(1); + expect(provider.scenarios).toHaveLength(2); expect(JSON.stringify(captured.structuredContent)).not.toContain( "correct horse battery staple", ); @@ -258,7 +260,7 @@ describe("browser scenario MCP tool", () => { arguments: scenario("https://other.example.test"), }); expect(denied.isError).toBe(true); - expect(provider.scenarios).toHaveLength(1); + expect(provider.scenarios).toHaveLength(2); const unapprovedEnvironment = scenario(); const secretDeclaration = unapprovedEnvironment.secrets.at(0); @@ -270,6 +272,6 @@ describe("browser scenario MCP tool", () => { arguments: unapprovedEnvironment, }); expect(deniedSecret.isError).toBe(true); - expect(provider.scenarios).toHaveLength(1); + expect(provider.scenarios).toHaveLength(2); }); }); diff --git a/tests/boundary/mcp/electronMcp.test.ts b/tests/boundary/mcp/electronMcp.test.ts index c4ab85c46f..afddfc78f7 100644 --- a/tests/boundary/mcp/electronMcp.test.ts +++ b/tests/boundary/mcp/electronMcp.test.ts @@ -37,7 +37,7 @@ afterEach(async () => { ); }); -it("exposes root-confined Electron discovery and inspection as Evidence v2", async () => { +it("exposes root-confined Electron discovery and inspection as Evidence", async () => { const root = await createTestTempDirectory("rea-electron-mcp-"); temporary.push(root); await writeFile(join(root, "index.html"), ""); diff --git a/tests/boundary/mcp/enhancedTools.01-lists-the-complete-target-open.test.ts b/tests/boundary/mcp/enhancedTools.01-lists-the-complete-target-open.test.ts index a20acdcb47..7da2cf16fd 100644 --- a/tests/boundary/mcp/enhancedTools.01-lists-the-complete-target-open.test.ts +++ b/tests/boundary/mcp/enhancedTools.01-lists-the-complete-target-open.test.ts @@ -188,15 +188,14 @@ describe("enhanced MCP tools", () => { ).toEqual(ENHANCED_TOOL_CONTRACTS.map(({ name }) => name).sort()); }); - it("executes all thirteen tools through production registration", async () => { + it("executes all twelve tools through production registration", async () => { const client = await connect(); const calls = [ - ["swift_classes", { pattern: "Fixture" }], ["get_objc_classes", { pattern: "Fixture" }], ["get_objc_protocols", {}], ["batch_decompile", { addresses: ["0x1", "0x2"] }], ["get_call_graph", { address: "0x1", direction: "forward" }], - ["analyze_swift_types", {}], + ["analyze_swift_types", { category: "classes", pattern: "Fixture" }], ["find_xrefs_to_name", { name: "entry" }], ["binary_overview", {}], ["analyze_function", { procedure: "0x1" }], @@ -212,8 +211,7 @@ describe("enhanced MCP tools", () => { ); expect(results[0]).toMatchObject({ count: 1 }); expect(results[1]).toMatchObject({ count: 1 }); - expect(results[2]).toMatchObject({ count: 1 }); - expect(results[3]).toEqual({ + expect(results[2]).toEqual({ items: [ { address: "0x1", status: "ok", pseudocode: "pseudo:0x1" }, { address: "0x2", status: "ok", pseudocode: "pseudo:0x2" }, @@ -222,31 +220,39 @@ describe("enhanced MCP tools", () => { succeeded: 2, failed: 0, }); - expect(results[4]).toEqual({ + expect(results[3]).toEqual({ "0": [{ address: "0x1", status: "ok", calls: ["0x2", "0x3"] }], "1": [ { address: "0x2", status: "ok", calls: ["0x1"] }, { address: "0x3", status: "ok", calls: [] }, ], }); - expect(results[5]).toMatchObject({ total: 6 }); - expect(results[6]).toEqual({ + expect(results[4]).toMatchObject({ + total: 1, + categories: { + classes: { + count: 1, + items: [{ address: "0x1", name: "_TtC7Fixture5Class" }], + }, + }, + }); + expect(results[5]).toEqual({ status: "resolved", name: "entry", address: "0x13", xrefs: ["0x20", "0x21"], }); - expect(results[7]).toMatchObject({ + expect(results[6]).toMatchObject({ document: "fixture", segment_count: 1, procedure_count: 6, string_count: 1, }); - expect(results[8]).toMatchObject({ + expect(results[7]).toMatchObject({ procedure: { address: "0x1", name: "entry" }, pseudocode: "return 0;", }); - expect(results[9]).toMatchObject({ + expect(results[8]).toMatchObject({ procedure: { address: "0x1", name: "entry" }, boundary: { available: false }, unsupported_branches: [ @@ -262,7 +268,7 @@ describe("enhanced MCP tools", () => { { operation: "preserve_residual_unknowns", status: "completed" }, ], }); - expect(results[10]).toMatchObject({ + expect(results[9]).toMatchObject({ query: "hello", search_mode: "literal", truncated: false, @@ -271,12 +277,12 @@ describe("enhanced MCP tools", () => { { target_address: "0x30", source_address: "0x21" }, ], }); - expect(results[11]).toMatchObject({ + expect(results[10]).toMatchObject({ query: "hello", matches: [{ type: "string", address: "0x30", value: "hello" }], truncated: false, }); - expect(results[12]).toMatchObject({ + expect(results[11]).toMatchObject({ start: "0x1", goal: "0x2", direction: "forward", diff --git a/tests/boundary/mcp/enhancedTools.03-returns-typed-graph-failures-and.test.ts b/tests/boundary/mcp/enhancedTools.03-returns-typed-graph-failures-and.test.ts index c4c06ef93a..fcff094e52 100644 --- a/tests/boundary/mcp/enhancedTools.03-returns-typed-graph-failures-and.test.ts +++ b/tests/boundary/mcp/enhancedTools.03-returns-typed-graph-failures-and.test.ts @@ -278,10 +278,13 @@ describe("enhanced MCP tools", () => { }, }); const result = jsonResult( - await client.callTool({ name: "swift_classes", arguments: {} }), + await client.callTool({ name: "analyze_swift_types", arguments: {} }), ); expect(calls).toEqual(["list_procedures"]); - expect(result).toMatchObject({ count: 2 }); + expect(result).toMatchObject({ + total: 2, + categories: { classes: { count: 2 } }, + }); }); it("returns cancellation when a complete inventory call is cancelled", async () => { @@ -296,7 +299,7 @@ describe("enhanced MCP tools", () => { }; const result = await new EnhancedTools(analysis).execute( - "swift_classes", + "analyze_swift_types", {}, controller.signal, ); diff --git a/tests/boundary/mcp/enhancedTools.04-uses-complete-inventories-inline.test.ts b/tests/boundary/mcp/enhancedTools.04-uses-complete-inventories-inline.test.ts index 9adb4934cc..99cefdaae3 100644 --- a/tests/boundary/mcp/enhancedTools.04-uses-complete-inventories-inline.test.ts +++ b/tests/boundary/mcp/enhancedTools.04-uses-complete-inventories-inline.test.ts @@ -269,7 +269,7 @@ describe("enhanced MCP tools", () => { ), }); const result = await client.callTool({ - name: "swift_classes", + name: "analyze_swift_types", arguments: {}, }); expect(result.isError).toBe(true); diff --git a/tests/boundary/mcp/enhancedTools.06-rejects-incomplete-function-dossiers-at.test.ts b/tests/boundary/mcp/enhancedTools.06-rejects-incomplete-function-dossiers-at.test.ts index 5657b8ab5d..9935e16f70 100644 --- a/tests/boundary/mcp/enhancedTools.06-rejects-incomplete-function-dossiers-at.test.ts +++ b/tests/boundary/mcp/enhancedTools.06-rejects-incomplete-function-dossiers-at.test.ts @@ -180,30 +180,4 @@ describe("enhanced MCP tools", () => { JSON.stringify(result.structuredContent), ); }); - - it("rejects legacy paged collection wrappers", async () => { - const malformedPort = fixturePort(); - const client = await connect({ - execute: async (name, arguments_, options) => { - const result = await malformedPort.execute(name, arguments_, options); - if (!result.ok || name !== "analyze_function") return result; - const dossier = result.value.result; - if ( - typeof dossier !== "object" || - dossier === null || - Array.isArray(dossier) - ) - return result; - return ok({ - ...dossier, - callers: { items: [], total: 1, returned: 0 }, - }); - }, - }); - const result = await client.callTool({ - name: "analyze_function", - arguments: { procedure: "0x1" }, - }); - expect(result.isError).toBe(true); - }); }); diff --git a/tests/boundary/mcp/enhancedTools.inputValidation.test.ts b/tests/boundary/mcp/enhancedTools.inputValidation.test.ts index 3768aead99..2f2724ac04 100644 --- a/tests/boundary/mcp/enhancedTools.inputValidation.test.ts +++ b/tests/boundary/mcp/enhancedTools.inputValidation.test.ts @@ -177,7 +177,7 @@ describe("enhanced MCP input validation", () => { execute: () => Promise.resolve(ok(["not", "a", "procedure", "map"])), }); const result = await client.callTool({ - name: "swift_classes", + name: "analyze_swift_types", arguments: {}, }); expect(result.isError).toBe(true); diff --git a/tests/boundary/mcp/functionComparisonMcp.test.ts b/tests/boundary/mcp/functionComparisonMcp.test.ts index 0322a6eab0..10cdea43d1 100644 --- a/tests/boundary/mcp/functionComparisonMcp.test.ts +++ b/tests/boundary/mcp/functionComparisonMcp.test.ts @@ -8,7 +8,7 @@ import { createServer } from "../../../src/server/createServer.js"; import { observed } from "../../fixtures/analysisExecution.js"; describe("function comparison MCP integration", () => { - it("records linked comparison Evidence and an approved residual unknown", async () => { + it("records linked comparison Evidence and its residual unknown", async () => { const session = createTestBinarySession(() => ({ health: () => Promise.resolve(), execute: () => Promise.resolve(observed(null)), @@ -35,7 +35,6 @@ describe("function comparison MCP integration", () => { arguments: { left_evidence_id: FUNCTION_COMPARISON_EXAMPLE.left.evidence_id, right_evidence_id: FUNCTION_COMPARISON_EXAMPLE.right.evidence_id, - unknown_registry_approved: true, }, }); expect(result.isError).not.toBe(true); diff --git a/tests/boundary/mcp/ghidraRoutingMcp.test.ts b/tests/boundary/mcp/ghidraRoutingMcp.test.ts index 7153d9345e..6bcfeafabd 100644 --- a/tests/boundary/mcp/ghidraRoutingMcp.test.ts +++ b/tests/boundary/mcp/ghidraRoutingMcp.test.ts @@ -40,7 +40,6 @@ describe("Ghidra MCP capability routing", () => { name: "binary_overview", available: true, }), - expect.objectContaining({ name: "swift_classes", available: true }), expect.objectContaining({ name: "get_objc_classes", available: true, diff --git a/tests/boundary/mcp/investigationAnalysisMcp.test.ts b/tests/boundary/mcp/investigationAnalysisMcp.test.ts index a48b6cf9fd..65a6569348 100644 --- a/tests/boundary/mcp/investigationAnalysisMcp.test.ts +++ b/tests/boundary/mcp/investigationAnalysisMcp.test.ts @@ -38,7 +38,6 @@ it("aggregates comparison Evidence and records an approved runtime gap", async ( name: "find_changed_behavior", arguments: { comparisons: [comparison], - unknown_registry_approved: true, }, }); expect(response.isError, JSON.stringify(response)).not.toBe(true); @@ -123,7 +122,6 @@ it("records a safe unresolved call-path question", async () => { functions: [dossier], start: { address: "0x1000" }, goal: { address: "0x3000" }, - unknown_registry_approved: true, }, }); expect(response.isError, JSON.stringify(response)).not.toBe(true); @@ -197,7 +195,6 @@ it("uses inline assembly observations without recording a false unknown", async }, }, ], - unknown_registry_approved: true, }, }); expect(response.isError).not.toBe(true); @@ -222,7 +219,6 @@ it("passes only the finite declared reconstruction specification", async () => { name: "verify_reconstruction", arguments: { ...INVESTIGATION_EXAMPLES.verify_reconstruction, - unknown_registry_approved: true, }, }); expect(response.isError).not.toBe(true); @@ -271,7 +267,6 @@ it("cannot omit a session-owned active unknown from reconstruction input", async name: "verify_reconstruction", arguments: { ...INVESTIGATION_EXAMPLES.verify_reconstruction, - unknown_registry_approved: true, }, }); expect(response.isError).not.toBe(true); diff --git a/tests/boundary/mcp/processTraceAdaptersParity.test.ts b/tests/boundary/mcp/processTraceAdaptersParity.test.ts index 10ea544e62..ab5fb79516 100644 --- a/tests/boundary/mcp/processTraceAdaptersParity.test.ts +++ b/tests/boundary/mcp/processTraceAdaptersParity.test.ts @@ -42,7 +42,6 @@ const capture = { ], } as const; const traceSpecification: ProcessTraceSpecification = { - version: 1, events: [ { id: "before", diff --git a/tests/boundary/mcp/serverIdentity.test.ts b/tests/boundary/mcp/serverIdentity.test.ts index 1b782e6bf5..f6a711fa50 100644 --- a/tests/boundary/mcp/serverIdentity.test.ts +++ b/tests/boundary/mcp/serverIdentity.test.ts @@ -60,8 +60,6 @@ const statusCapability = ( } = { available: true }, ) => ({ operation, - input_contract_version: 1, - output_contract_version: 1, effects: { mutates_artifact: false, launches_process: false, diff --git a/tests/boundary/mcp/serverUnknownsMcp.test.ts b/tests/boundary/mcp/serverUnknownsMcp.test.ts index 0addc58309..e10198ae97 100644 --- a/tests/boundary/mcp/serverUnknownsMcp.test.ts +++ b/tests/boundary/mcp/serverUnknownsMcp.test.ts @@ -75,7 +75,7 @@ const structured = (result: CallToolResult): Record => { return Object.fromEntries(Object.entries(result.structuredContent)); }; -it("records approved typed capability unavailability without forwarding the flag", async () => { +it("does not record capability unavailability without supporting Evidence", async () => { const received: Array>> = []; const analysis: AnalysisOperationPort = { execute: (name, arguments_) => { @@ -110,7 +110,6 @@ it("records approved typed capability unavailability without forwarding the flag name: "procedure_pseudo_code", arguments: { procedure: "main", - unknown_registry_approved: true, }, }); expect(unavailable.isError).toBe(true); @@ -119,24 +118,12 @@ it("records approved typed capability unavailability without forwarding the flag category: "unsupported_provider", }, }); - expect(received[0]).not.toHaveProperty("unknown_registry_approved"); + expect(received[0]).toMatchObject({ procedure: "main" }); expect( structured(await client.callTool({ name: "list_unknowns", arguments: {} })), ).toMatchObject({ result: { - items: [ - { - domain: "analysis-capability", - question: - "The requested analysis is unavailable for the current target.", - recommended_probes: [ - { - rationale: - "Choose another analysis or target that can answer this question.", - }, - ], - }, - ], + items: [], }, }); }); @@ -214,7 +201,6 @@ it("records approved capture disagreement as a contradicted unknown", async () = arguments: { left_evidence_id: leftEvidence.evidence_id, right_evidence_id: rightEvidence.evidence_id, - unknown_registry_approved: true, }, }); expect(compared.isError).not.toBe(true); diff --git a/tests/boundary/mcp/serverWorkflowMcp.test.ts b/tests/boundary/mcp/serverWorkflowMcp.test.ts index 5a1fda416f..48b2afb158 100644 --- a/tests/boundary/mcp/serverWorkflowMcp.test.ts +++ b/tests/boundary/mcp/serverWorkflowMcp.test.ts @@ -43,6 +43,32 @@ const structured = (result: CallToolResult): Record => { return Object.fromEntries(Object.entries(result.structuredContent)); }; +it("advertises the native function discovery-to-dossier workflow", async () => { + const client = await connect({ + execute: () => Promise.resolve(ok(null)), + }); + + expect(client.getInstructions()).toContain( + "call binary_overview for target metadata and inventory counts, then search_procedures or list_procedures to choose a function", + ); + + const tools = new Map( + (await client.listTools()).tools.map((tool) => [tool.name, tool]), + ); + expect(tools.get("binary_overview")?.description).toContain( + "follow with search_procedures or list_procedures, then pass a returned name or address to analyze_function", + ); + expect(tools.get("analyze_function")?.description).toContain( + "typically selected from search_procedures or list_procedures", + ); + expect(tools.get("analyze_function")?.inputSchema.properties).toHaveProperty( + "procedure", + ); + expect(tools.get("inspect_native_api")?.description).toContain( + "use analyze_function for a general function dossier", + ); +}); + it("executes a realistic workflow: list methods, decompile selected, get xrefs", async () => { const client = await connect({ execute: (name, args) => { diff --git a/tests/boundary/mcp/sessionMcp.test.ts b/tests/boundary/mcp/sessionMcp.test.ts index b39523356d..e2eec91a1c 100644 --- a/tests/boundary/mcp/sessionMcp.test.ts +++ b/tests/boundary/mcp/sessionMcp.test.ts @@ -229,7 +229,7 @@ describe("target-free MCP workflow", () => { }); describe("process residuals over MCP", () => { - it("records approved process residuals in the unknown registry", async () => { + it("records process residuals linked to capture Evidence", async () => { if (!(await probeProcessCaptureCapability()).available) return; const session = createTestBinarySession(() => client("fixture", [])); const server = createServer(session, session, { @@ -252,7 +252,6 @@ describe("process residuals over MCP", () => { const captured = await mcp.callTool({ name: "capture_process_scenario", arguments: { - unknown_registry_approved: true, executable: process.execPath, arguments: [processFixture, "partial"], working_directory: dirname(processFixture),