Commit Graph
42 Commits
Author SHA1 Message Date
morluto b51ed4c3ec refactor(kernel): migrate the remaining SHA-256 hex digests to sha256Hex and move isRecord into src/kernel (#11571)
* refactor(kernel): migrate the remaining one-shot SHA-256 hex digests to sha256Hex

* refactor(kernel): move isRecord into src/kernel and replace its inline copies (#11574)

* fix(scripts): import isRecord from src/kernel in the managed manifest parser
2026-11-15 20:53:36 -05:00
morluto 5f718ce30e refactor(kernel): give error-message, byte-hex and SHA-256 hex helpers one owner in a new src/kernel layer (#11542)
* refactor(kernel): give error-to-message text one owner in a new src/kernel layer

* fix(kernel): keep the regex worker's own error text inside its isolated source

* refactor(kernel): exclude copies that run in another realm from the error-message guard by name

The regex worker's source string and the page.evaluate fetch in PlaywrightScenarioRequest cannot import errorMessage. The guard now names them, so the counted baseline does not invite a later slice to rewrite them.

* refactor(kernel): migrate the long tail of inline error-message copies to errorMessage (#11552)

Every provider and adapter folder with six or fewer copies now calls errorMessage from src/kernel. Local helpers whose whole body was the copy (failureMessage, gdbErrorMessage, message, describe, reasonOf, recoveryFailureMessage and four local errorMessage definitions) are removed and their call sites call the owner.

Refs #11548

* refactor(kernel): give byte-to-hex formatting one owner in src/kernel (#11560)

byteToHex returns the two lowercase hexadecimal digits of one byte. All eight inline copies (native, contracts, domain firmware and desktop, pcap) call it, keeping their own 0x, % or \x notation, and a counted guard with an empty baseline keeps new copies out.

Refs #11548

* refactor(kernel): give one-shot SHA-256 hex digests one owner in src/kernel (#11562)

sha256Hex digests one in-memory value. A counted guard records the one-shot createHash("sha256").update(x).digest("hex") chains (115 in 101 files) and the first slice migrates domain, contracts and process (35 copies), removing local wrappers (sha256, sha256Text, hash) and the exported sha256Text. Incremental and streamed hashing is not counted.

Refs #11548

* refactor(kernel): migrate the application layer's inline error-message copies to errorMessage (#11563)

All 69 copies in src/application call errorMessage from src/kernel. Local helpers whose whole body was the copy (errorDetail, causeReason and five local errorMessage definitions) are removed and their call sites call the owner.

Refs #11548

* refactor(kernel): migrate the native provider's inline error-message copies to errorMessage (#11567)

All 49 copies in src/native call errorMessage from src/kernel. The local errorReason helpers in LldbCallTracer and LldbRetainedObservations, whose whole body was the copy, are removed and their call sites call the owner.

Refs #11548

* refactor(kernel): migrate the remaining inline error-message copies to errorMessage (#11569)
2026-11-16 07:05:30 +08:00
morluto 960a387731 refactor(contracts): own CLI command names, catalog identity and schema presentation in contracts (#11425)
* refactor(contracts): move the CLI command name owner into src/contracts

* refactor(contracts): own catalog identity in contracts and server identity in process (#11435)

* refactor(contracts): move catalog identity into contracts and server identity into process

* refactor(contracts): own the advertised MCP schema presentation in contracts

* test(mcp): import the advertised presentation from contracts in toolSchemaValidity

* test(release): expect the catalog identity at its contracts path in the canary

* test(mcp): import the catalog identity from contracts in syntheticContractSurfaces

* refactor(contracts): import CLI command names from contracts in the capability conformance test
2026-11-16 05:31:37 +08:00
morluto 47c252f527 refactor(config): move the MCP startup policy beside its startup budgets in src/config (#11429) 2026-11-16 04:14:55 +08:00
morluto 5c53bcaed8 feat(mcp)!: retire the full schema profiles and declare schema facts as typed annotations (#10693)
* test(mcp): check advertised schemas against a portable dialect in every profile

* fix(mcp): re-emit advertised patterns in a portable regex grammar

* fix(mcp): present discover_tools input through the selected profile

The hot-cold catalog registered discover_tools directly with the SDK, so
the default compact input profile still advertised its canonical schema,
including page-size defaults clients apply on their own (#5322). Route
every listed input schema, transport discovery included, through one
server-owned presentation, and pass the selected presentation to the
deferred server, which serves recorded tool schemas but not discovery.

* fix(mcp): keep the Evidence union on described borrowed inputs

borrowEvidenceInput installed its advertised projection on one schema
instance. A call site's .describe() rebuilds the schema from its
definition, so match_native_functions advertised left and right as
unconstrained values in every input profile. Install the projection in
the borrowed parser's constructor so derived schemas keep it, for both
borrowed inputs and borrowed arrays.

* fix(mcp): validate advertised dialect by selected profile

* fix(mcp): preserve schema widening through references

* feat(mcp): project compact input schemas into the portable advertised dialect

One projection (advertisedDialectProjection.ts) owns the compact input
advertisement: one type per node, items on every array, string literals
as enums, oneOf widened to anyOf, allOf merged, tuples as element unions,
exclusive bounds made inclusive, and every keyword outside the dialect
dropped. Each rewrite is exact or widening; server validation is
unchanged. The root description and sentences for dropped cross-field
constraints move to the tool description, which Codex and Zed keep.

The compact baseline falls to nullable values only (29), which the
dialect cannot express without narrowing an input. The gateway shape
repairs (typeArrayItemUnions, hasGatewayIncompatibleShape), the second
dependentRequired-to-prose copy and the strict-gateway test denylist are
removed; the dialect checker gates the inline presentation instead.

* feat(mcp): omit output schemas from tools/list by default

No surveyed model API reads an output schema, MCP client validators
repeat REA's own result validation, and complete output schemas were
about 92% of the default tools/list. REA_MCP_OUTPUT_SCHEMA_PROFILE now
resolves to off through one owner (resolveMcpOutputSchemaProfile);
compact and full remain selectable. structuredContent is still
delivered, and the deferred server serves both off and full from the
recorded catalog with the matching recorded identity.

Tests and artifact verifiers that validate results against complete
advertised output contracts select full explicitly.

* docs: describe omitted output schemas as the default in README translations

* fix(mcp): select schemas in verification workflows

* fix(mcp): refuse zero-padded repeat bounds and record host-specific path patterns

* fix(mcp): keep projected schemas permissive

* fix(mcp): keep references valid when presentation widens their targets

* fix(mcp): keep a single union branch's closure scoped to its properties

* fix(mcp): select complete output schemas in the Electron verifier

* fix(mcp): select complete output schemas in standalone verifier probes

* fix(mcp): keep an allOf member's tuple tail with its own positions

* fix(mcp): relax references into negations and conditionals presentation may remove

* refactor(contracts): declare root discriminators as typed schema annotations

* test(mcp): select the borrowed Evidence contract from its family

* fix(mcp): keep astral literals and dynamic references widening in every mode

* fix(mcp): omit astral literals and relax untracked references when presentation widens

* fix(mcp): select complete output schemas in every verifier transport

* fix(mcp): read reference pointers by schema position, not segment spelling

* fix(mcp): protect references into collapsing boolean unions and keep referencing branches out of the collapse

* fix(mcp): keep reduced schemas open when their root admits names by pattern

* test(mcp): keep the reduced-schema pattern case outside the size-limited block

* fix(mcp): omit patterns whose classes name astral characters

* fix(mcp): decide union collapse and pointer protection from canonical branches

* docs(mcp): document non-u astral matching and pin BMP agreement across V8 modes

* fix(mcp): keep prototype-named properties as own properties in the dialect projection

* test(mcp): select full output schemas in packet and signal schema checks

* test(mcp): expect the advertised base64 schema to omit the unbounded pad-bit probe

* fix(mcp): protect references into unevaluated keywords lost evaluation removes

* fix(mcp): validate the doctor probe reply against the canonical output contract (#10655)

* fix(mcp): validate the doctor probe reply against the canonical output contract

* test(mcp): replay doctor probe validation under every output profile

* test(mcp): accept typed-branch nullable unions in the advertised dialect (#10649)

* test(mcp): select full output schemas in HTTP session workflows

* fix(mcp): emit stack-safe classes with non-ASCII members and protect draft-07 reference siblings

* fix(mcp): protect references into literal siblings of relaxed draft-07 references

* feat(mcp): omit output schemas from tools/list by default (#10646)

* feat(mcp): omit output schemas from tools/list by default

No surveyed model API reads an output schema, MCP client validators
repeat REA's own result validation, and complete output schemas were
about 92% of the default tools/list. REA_MCP_OUTPUT_SCHEMA_PROFILE now
resolves to off through one owner (resolveMcpOutputSchemaProfile);
compact and full remain selectable. structuredContent is still
delivered, and the deferred server serves both off and full from the
recorded catalog with the matching recorded identity.

Tests and artifact verifiers that validate results against complete
advertised output contracts select full explicitly.

* fix(mcp): select schemas in verification workflows

* fix(mcp): select complete output schemas in the Electron verifier

* fix(mcp): select complete output schemas in standalone verifier probes

* fix(mcp): select complete output schemas in every verifier transport

* test(mcp): select full output schemas in packet and signal schema checks

* fix(mcp): validate the doctor probe reply against the canonical output contract (#10655)

* fix(mcp): validate the doctor probe reply against the canonical output contract

* test(mcp): replay doctor probe validation under every output profile

* test(mcp): accept typed-branch nullable unions in the advertised dialect (#10649)

* test(mcp): select full output schemas in HTTP session workflows

* fix(mcp): keep the Evidence union on described borrowed inputs (#10597)

* fix(mcp): keep the Evidence union on described borrowed inputs

borrowEvidenceInput installed its advertised projection on one schema
instance. A call site's .describe() rebuilds the schema from its
definition, so match_native_functions advertised left and right as
unconstrained values in every input profile. Install the projection in
the borrowed parser's constructor so derived schemas keep it, for both
borrowed inputs and borrowed arrays.

* test(mcp): select the borrowed Evidence contract from its family

* docs(mcp): admit baseline growth only for restored contract declarations

* fix(mcp): state reduced multi-type properties as typed branches, not type lists

The reduced fallback derived a property's types and emitted them as a type
list, which Gemini, Moonshot and Ark reject. A property whose canonical
schema is anyOf string and number therefore traded a typed union for a
rejected construct when the inline projection fell back. Emit one type, or
anyOf typed branches, so [T, null] becomes the dialect's nullable union.

* feat(mcp)!: retire the full schema profiles and declare complete observations as a typed annotation (#10718)

* refactor(mcp): declare complete observations as a typed annotation applied only on the wire

* test(mcp): compare complete-observation declarations with their owning text

* refactor(schema): couple Zod checks with their JSON Schema constraint keywords (#10725)

* refactor(schema): couple Zod checks with their JSON Schema constraint keywords

* refactor(schema): declare pattern approximations and object output roots as typed annotations (#10754)

* refactor(schema): declare pattern approximations and object output roots as typed annotations

* fix(schema): refuse pattern declarations JSON Schema cannot state

* test(schema): select guided_workflows from its contract family

* feat(mcp)!: retire the full input and output schema profiles (#10841)

* feat(mcp)!: retire the full input and output schema profiles

The compact input presentation is now the only advertised input schema, and
output schemas are either omitted (`off`, the default) or bounded
(`compact`). Catalog identity, the build-recorded digests and the recorded
startup catalog collapse to those presentations: two digest profiles instead
of six, one recorded catalog instead of two, and no input dialect baseline.

Tests and verifiers that validated replies against listed full output
schemas now validate against the canonical output contract, which REA
already enforces server-side; tests that asserted canonical input
restrictions read the canonical input schema instead of a full listing.

BREAKING CHANGE: `REA_MCP_INPUT_SCHEMA_PROFILE=full` and
`REA_MCP_OUTPUT_SCHEMA_PROFILE=full` are refused at startup and by `rea setup`
with a configuration error naming the setting. Remove the setting from the
client registration (or rerun setup with `REA_MCP_INPUT_SCHEMA_PROFILE=compact`
/ `REA_MCP_OUTPUT_SCHEMA_PROFILE=compact`). `REA_MCP_INPUT_SCHEMA_PROFILE`
still accepts `compact`. Server-side validation is unchanged.

* test(mcp): validate replies against canonical output contracts at every listed-schema reader

tools/list omits output schemas by default, so fixtures, verifiers and the
Android analysis runners that read a listed outputSchema now bind the
canonical contract instead. connectLocalToolsMcp and advertisedMcpToolSchemas
hand the SDK a definition carrying the canonical output schema, so the SDK
client and caller-supplied validators check every successful reply.

* test(mcp): resolve Android runners' output contracts from the selected repository

The Android efficiency runners imported the canonical output-schema helper from the harness checkout while --repository selects the revision under test. Import it from the selected repository, as the runners already do for the SDK and Ajv.

* test(mcp): let Android runners compare revisions before the canonical output helper

A selected revision without scripts/lib/canonical-output-schema.mjs could not start the Android runners. Choose the output contracts in one harness helper: the selected revision's canonical contracts when it has them, otherwise its listed schemas under the full output profile it still supports.

* fix(scripts): keep caller launch settings when the real Hopper verifier builds its server environment

The verifier deleted REA_VERIFY_SERVER_COMMAND and REA_VERIFY_SERVER_ARGS_JSON from process.env itself before reading them, so a caller selecting an installed REA command silently ran the checkout's dist/main.js. Copy the environment, as main does, so only the child's environment changes.

* test(mcp): validate whole replies against canonical output contracts wherever a listed definition is held

tools/list omits output schemas by default and the full output profile is gone, so tests that handed the SDK a listed definition no longer checked complete replies. Attach the canonical output schema with withCanonicalOutputSchema at every held REA tool definition. Environment-served tools (not REA contracts), and the deferred-runtime and doctor tests that exercise per-profile listings, keep their listed definitions.

* test(mcp): record the advertised catalog summary without the retired full profiles

Retiring the full input and output profiles leaves only the compact input and compact output listings, so the summary drops the full-profile digests and every pattern only the full listings carried. Regenerated with npm run verify:advertised-schemas -- --update-summary.
2026-11-15 21:51:18 +08:00
morluto 4126698407 feat(mcp): project compact input schemas into the portable advertised dialect (#10632)
* test(mcp): check advertised schemas against a portable dialect in every profile

* fix(mcp): re-emit advertised patterns in a portable regex grammar

* fix(mcp): present discover_tools input through the selected profile

The hot-cold catalog registered discover_tools directly with the SDK, so
the default compact input profile still advertised its canonical schema,
including page-size defaults clients apply on their own (#5322). Route
every listed input schema, transport discovery included, through one
server-owned presentation, and pass the selected presentation to the
deferred server, which serves recorded tool schemas but not discovery.

* fix(mcp): keep the Evidence union on described borrowed inputs

borrowEvidenceInput installed its advertised projection on one schema
instance. A call site's .describe() rebuilds the schema from its
definition, so match_native_functions advertised left and right as
unconstrained values in every input profile. Install the projection in
the borrowed parser's constructor so derived schemas keep it, for both
borrowed inputs and borrowed arrays.

* fix(mcp): validate advertised dialect by selected profile

* fix(mcp): preserve schema widening through references

* feat(mcp): project compact input schemas into the portable advertised dialect

One projection (advertisedDialectProjection.ts) owns the compact input
advertisement: one type per node, items on every array, string literals
as enums, oneOf widened to anyOf, allOf merged, tuples as element unions,
exclusive bounds made inclusive, and every keyword outside the dialect
dropped. Each rewrite is exact or widening; server validation is
unchanged. The root description and sentences for dropped cross-field
constraints move to the tool description, which Codex and Zed keep.

The compact baseline falls to nullable values only (29), which the
dialect cannot express without narrowing an input. The gateway shape
repairs (typeArrayItemUnions, hasGatewayIncompatibleShape), the second
dependentRequired-to-prose copy and the strict-gateway test denylist are
removed; the dialect checker gates the inline presentation instead.

* fix(mcp): refuse zero-padded repeat bounds and record host-specific path patterns

* fix(mcp): keep projected schemas permissive

* fix(mcp): keep references valid when presentation widens their targets

* fix(mcp): keep a single union branch's closure scoped to its properties

* fix(mcp): keep an allOf member's tuple tail with its own positions

* fix(mcp): relax references into negations and conditionals presentation may remove

* test(mcp): select the borrowed Evidence contract from its family

* fix(mcp): keep astral literals and dynamic references widening in every mode

* fix(mcp): omit astral literals and relax untracked references when presentation widens

* fix(mcp): read reference pointers by schema position, not segment spelling

* fix(mcp): protect references into collapsing boolean unions and keep referencing branches out of the collapse

* fix(mcp): keep reduced schemas open when their root admits names by pattern

* test(mcp): keep the reduced-schema pattern case outside the size-limited block

* fix(mcp): omit patterns whose classes name astral characters

* fix(mcp): decide union collapse and pointer protection from canonical branches

* docs(mcp): document non-u astral matching and pin BMP agreement across V8 modes

* fix(mcp): keep prototype-named properties as own properties in the dialect projection

* test(mcp): expect the advertised base64 schema to omit the unbounded pad-bit probe

* fix(mcp): protect references into unevaluated keywords lost evaluation removes

* fix(mcp): emit stack-safe classes with non-ASCII members and protect draft-07 reference siblings

* fix(mcp): protect references into literal siblings of relaxed draft-07 references

* feat(mcp): omit output schemas from tools/list by default (#10646)

* feat(mcp): omit output schemas from tools/list by default

No surveyed model API reads an output schema, MCP client validators
repeat REA's own result validation, and complete output schemas were
about 92% of the default tools/list. REA_MCP_OUTPUT_SCHEMA_PROFILE now
resolves to off through one owner (resolveMcpOutputSchemaProfile);
compact and full remain selectable. structuredContent is still
delivered, and the deferred server serves both off and full from the
recorded catalog with the matching recorded identity.

Tests and artifact verifiers that validate results against complete
advertised output contracts select full explicitly.

* fix(mcp): select schemas in verification workflows

* fix(mcp): select complete output schemas in the Electron verifier

* fix(mcp): select complete output schemas in standalone verifier probes

* fix(mcp): select complete output schemas in every verifier transport

* test(mcp): select full output schemas in packet and signal schema checks

* fix(mcp): validate the doctor probe reply against the canonical output contract (#10655)

* fix(mcp): validate the doctor probe reply against the canonical output contract

* test(mcp): replay doctor probe validation under every output profile

* test(mcp): accept typed-branch nullable unions in the advertised dialect (#10649)

* test(mcp): select full output schemas in HTTP session workflows

* fix(mcp): keep the Evidence union on described borrowed inputs (#10597)

* fix(mcp): keep the Evidence union on described borrowed inputs

borrowEvidenceInput installed its advertised projection on one schema
instance. A call site's .describe() rebuilds the schema from its
definition, so match_native_functions advertised left and right as
unconstrained values in every input profile. Install the projection in
the borrowed parser's constructor so derived schemas keep it, for both
borrowed inputs and borrowed arrays.

* test(mcp): select the borrowed Evidence contract from its family

* docs(mcp): admit baseline growth only for restored contract declarations

* fix(mcp): state reduced multi-type properties as typed branches, not type lists

The reduced fallback derived a property's types and emitted them as a type
list, which Gemini, Moonshot and Ark reject. A property whose canonical
schema is anyOf string and number therefore traded a typed union for a
rejected construct when the inline projection fell back. Emit one type, or
anyOf typed branches, so [T, null] becomes the dialect's nullable union.
2026-11-15 18:26:13 +08:00
morluto 5bf66f7500 refactor(domain): one owner for the non-array record predicate (#10588)
* refactor(domain): one owner for the non-array record predicate

Thirty-four source files each declared isRecord as a non-null object that is
not an array, and six more spelled the same predicate as isObject,
isSchemaRecord or isStorageRecord. They now import isRecord from
src/domain/isRecord.ts, whose overloads keep JsonValue narrowing for the
desktop and environment readers. DshCordisPatch's predicate also excludes
Cordis expressions, so it is renamed isPatchRecord.

architecture-guards refuses a new local copy under any name. The seven files
that still hold one are listed and must leave the list when their copy goes.

* fix(guards): count record predicate copies precisely

* test(hopper): retry after confirmed application exit

* fix: reject stale record predicate exceptions
2026-11-15 17:05:54 +08:00
morluto e68e548a3d fix(cli): validate plist and dylib-root options with their contract fields (#10557)
* fix(cli): validate plist and dylib-root options with their contract fields

inspect-plist --relative-path and trace-dylib-resolution --root were
declared as bare strings on the CLI. --schema under-advertised them, and an
empty or NUL value reached the operation contract, whose refusal named the
contract field (path, roots) instead of the CLI option.

Declare both options with the contract's own field schemas
(localPathStringSchema; dylibResolutionInputSchema.shape.roots), so the CLI
refuses them first and names --relative-path or --root.

Fixes #9764
Fixes #9823

* fix(mcp): observe SDK initiated cleanup failures
2026-11-14 14:50:59 -05:00
3a13b088e6 fix(mcp): explain skipped doctor identity probes
Report identity as unobserved when the selected tool families omit
binary_session, preserving the current health decision and exit status.
Cover scoped production sessions and retain validation of successful
and schema-invalid probe responses.

Fixes #7438

[gensokyo-dev: rebased onto main; report "identity not observed" for every
request-flow reply that was not received (skipped scope, transport failure or
rejected output schema), not only skipped scopes; corrected the replayed
inventory-mode expectation, whose identity is observed and matches.]

Fixes #7438

Co-Authored-By: gensokyo-dev <gensokyo-dev@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-14 05:36:32 +01:00
gensokyo-devandClaude Opus 5.5 d412569293 fix(mcp-doctor): close only through the owned transport
The owned transport is the doctor's only cleanup owner, and Client.close
only delegates to it. Close it once and report its typed failure instead
of aggregating a duplicate SDK close and clearing that failure after a
retry. Pass the filtered caller environment as the launcher's base only,
and name the doctor operation in transport cleanup errors.

Cover a descendant in its own session, which only the token sweep can
reach, and a late cleanup denial that keeps the startup diagnostic and
reports the retained run. Bound the fixture descendant's lifetime so a
failing run cannot leak it. Document the doctor's process-family
ownership and output bound.

Co-Authored-By: gensokyo-dev <gensokyo-dev@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-14 05:36:23 +01:00
2SinceJuly 1b019a73d8 fix(mcp-doctor): keep the diagnostic child target-free and its stderr visible
The owned launcher merges its host base back under the explicit overrides, so
filtering only the override object let an ambient HOPPER_TARGET_* variable
through and the diagnostic child could open the target the doctor promises to
skip. Pass the filtered environment as the host base too, and opt into the
token sweep so a detached descendant that outlives the launcher is still
reached by cleanup.

Also keep the stderr observer attached until bounded cleanup and draining have
settled, report the typed cleanup resources instead of a generic close_binary
message, and drop a first-attempt cleanup failure once the owned retry proves
release.
2026-11-14 05:36:23 +01:00
2SinceJuly d427fc05c3 fix(mcp-doctor): stop the owned process family on every exit path
runProductionMcpDoctor wrapped the SDK StdioClientTransport and only joined its
close calls. That transport clears its direct-child reference, ends stdin and
signals only that child, so a server whose launcher exited after forking an
inherited-stdio descendant left the descendant running while the doctor
reported nothing but the transport deadline.

Run the doctor on OwnedMcpStdioTransport, which spawns the child in a dedicated
POSIX process group or Windows job and verifies group and token-owned
descendant cleanup before close returns. Feed stderr through a sink on
acquisition so diagnostics survive an early child exit and still honor the
64 KiB retained prefix and full byte count.

Covered by a fixture that forks a descendant, exits, and is timed out: the
regression asserts the descendant is gone after the doctor returns.
2026-11-14 05:36:23 +01:00
morluto 5044e20171 test: focus diagnostics and evidence checks on behavior (#9303)
* test: focus configuration diagnostics on their owners

Preserve shape errors, selected paths, recovery guidance, and unchanged-file behavior across parser, setup, and uninstall boundaries.

* test: focus artifact and evidence checks on behavior

Retain format, identity, malformed-input, and cleanup coverage while removing fixed prose and closing the admitted test source.

* refactor(debugger): use configured session admission limits

* test: focus evidence and diagnostics on behavior

* fix(test): restore verifier export and formatting

* test(config): verify debugger limit ranges

* test(rizin): keep fixture executable authoritative

* test(rizin): isolate debugger fixture environment
2026-11-08 16:39:20 +00:00
morluto 352a846127 test: reduce brittle assertions and test-only runtime paths (#9289)
* test: focus assertions on behavioral evidence

Prune redundant and wording-coupled checks while retaining diagnostic, parser, recovery, and cleanup coverage.

* refactor: remove test-only wrappers and duplicate checks

Exercise production operations through test fixtures, remove unused adapters, and keep managed-member uncertainty guidance under its existing owner.

* test: focus process validation on diagnostic facts

* refactor: remove test-only runtime helpers

Move semantic evidence fixture setup out of runtime and remove unused Windows settle recovery helpers while checking configured defaults independently.

* test: focus artifact diagnostics on format facts

* test: verify retained cleanup guidance delivery
2026-11-08 02:02:31 +08:00
gensokyo-dev 0beedcd712 perf(mcp): add opt-in hot/cold tool discovery (#9046)
Add per-connection discovery that loads canonical named tools on demand while preserving the default full catalog. Bound discovery replies before changing visibility, preserve family exclusions and Evidence workflows, and reset visibility on HTTP reconnect.

Verified with focused protocol tests and real Codex CLI artifact investigations; token costs remain task-dependent, so the profile is opt-in.
2026-11-05 20:27:31 +00:00
Ha1baraA1 b3e22620a5 fix(mcp): keep scoped doctor output parseable (#7648) 2026-11-03 14:47:16 +08:00
morluto 60ab4da329 fix(cli): correct completion options and setup guidance (#7851)
* fix(cli): align MCP doctor completion with its options

* fix(cli): explain zsh completion initialization

* fix(cli): include launcher timing in doctor completion

* style(cli): format doctor completion schema
2026-10-29 01:28:04 +08:00
5ae77656e7 feat(workflows): expose guided prompts through CLI and MCP (#6995)
* feat(workflows): expose guided prompts through CLI and MCP

* fix(workflows): address guided workflow review findings

* fix(workflows): preserve scoped contracts and actionable validation

* test(workflows): allow bounded CLI startup on shared hosts

* test(workflows): audit guidance effects and content provenance

* perf(workflows): defer rendering and separate catalog scope

---------

Co-authored-by: Deepak-Kambala <deepakkambala07@gmail.com>
Co-authored-by: gensokyo-dev <342580529+gensokyo-dev@users.noreply.github.com>
2026-10-28 14:12:39 +08:00
aa6f6436c9 feat(cli): add CSV and TSV analysis views (#7254)
* feat(cli): add CSV and TSV analysis views

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Update tests/boundary/cli/filteredArrayOutput.test.ts

Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com>

* Update src/cli.ts

Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com>

* fix(cli): address delimited export review feedback

* fix(cli): format delimited output block in cliResultOutput

* fix(cli): retain delimited view uncertainty and align completions

---------

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com>
Co-authored-by: gensokyo-dev <342580529+gensokyo-dev@users.noreply.github.com>
2026-10-28 11:27:40 +08:00
b675073b73 feat(mcp): scope advertised tools by family (#7228)
Scope MCP discovery, setup, doctor, status and recovery guidance by selected tool families using one contract inventory and shared catalog identity. Preserve default discovery, canonical validation and complete Evidence.

Fixes #3246. Builds on #3026. Real libcrypto workflow verifies all 5,367 exports and identical complete Evidence. Static/source/build/docs/Windows/package/Greptile checks pass; ten assertions reproduced identically on clean main and feature across Node 24/22, with cancelled coverage explicitly unverified. Merge uses the user-authorized bypass for baseline failures and unchecked human-readiness declarations, without attesting human review.

Co-authored-by: 0xLLVEL <fakhrezan1@gmail.com>
Co-authored-by: Codebuff <noreply@codebuff.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-10-26 19:50:12 +08:00
morluto d69a54d592 fix(setup): register explicit MCP runtimes and preserve launcher identity 2026-10-26 02:33:31 +08:00
kiranandgensokyo-dev 5d2c8c2b78 feat(cli)!: export row-oriented results as JSON Lines (#5629)
* feat(cli): add JSON Lines output for row-oriented results (#5487)

- Support opt-in JSON Lines format (--format jsonl) for row-oriented results and collection views with clear record boundaries
- Preserve stable source identities and stream records deterministically line-by-line
- Emit final coverage and limitation metadata as a distinct line with https://rea.dev/schemas/jsonl-coverage-v1.json to prevent confusion with data rows
- Distinguish error outputs with https://rea.dev/schemas/jsonl-error-v1.json
- Reject non-row-oriented results that lack record boundaries with an unreadable_output error
- Document JSON Lines streaming conventions in docs/cli.md
- Add boundary tests covering empty results, multiline strings, partial coverage, large fields, and non-row rejection

* fix(cli): preserve diagnostics, provenance, and nested collections in JSONL

* fix(cli): frame classified failures and reject unrepresented JSONL facets

---------

Co-authored-by: gensokyo-dev <342580529+gensokyo-dev@users.noreply.github.com>
2026-10-24 02:13:28 +08:00
+12 169512a35b feat(cli): report version and runtime with an explicit output format (#6530)
* feat(cli): report version and runtime with an explicit output format

`rea --version --json` and `--version --format <format>` ignored the format
and printed only the bare version, so support reports gathered the runtime
and platform by hand.

A root --version request that selects a format now renders a documented
report: package name and version, JavaScript runtime and version, platform,
architecture, and whether the runtime satisfies the supported range. It
contains no paths, environment values, user or host names, and starts no
provider. Plain --version is unchanged, --help keeps precedence, and other
output controls keep Incur's plain version.

The runtime verdict uses a new shared supportsJavaScriptRuntime rule that
rea doctor now uses too, so both judge Bun by its Node.js compatibility
version. mcp doctor's private JSON/format renderer moves to cliOutput as
renderCliDocument and is shared with the version report.

Closes #5483

* refactor(errors): centralize input failure construction

Carry caller messages through the existing schema issue owner instead of constructing and discarding an intermediate input error. Remove forwarding schema helpers and migrate equivalent constructors across investigation, firmware, dylib, process, and screenshot workflows while preserving diagnostic paths and causes.

* docs(readme): add Roman Urdu edition (#6499)

* docs(readme): add Roman Urdu edition

Register ur-Latn with a pending website catalog and preserve English command examples.

Script-tag handling follows morluto's shared helper proposed in PR #2043 (c46f606c89).

* test(website): cover completed script-tag page metadata

* fix(process): parse native teardown relation sentinels

* fix(cli): make version report privacy checks environment-independent

The CLI test rejected any report containing the OS account name, so an
account named node, linux or rea failed a correct report. Check unique
caller-controlled HOME, cwd and environment markers instead; the strict
schema still rejects any additional field.

The CLI guide said --version with other output controls prints the plain
version, but invalid combinations such as --json with --token-limit report
their existing output-flag error first. Describe both outcomes and test them.

* refactor(nativeaot): share report facts across budget stages

Centralize parse-wide identity and coverage facts while keeping conservative admission values separate from measured rows and counters. Reuse the derived truncation reason and retain the final serialization bound.

Tighten the report-budget regression so unexpected parser finalization errors cannot be swallowed by its budget search.

* fix(adb): preserve empty properties and verbatim version names

* fix(artifacts): recognize empty Zip64 containers

Fixes #6540

* fix(erlang): accept uninterpreted BEAM alignment bytes

Fixes #6517

* fix(gpu): accept PTX entries without parameter lists

Fixes #6493

* fix(python): preserve signed code-object line numbers

Fixes #6490

* refactor(ghidra): centralize load image check recording

Share expected and observed check construction across MZ, COM, and raw-binary attestations. Preserve format-specific rules and source coordinates, and derive mapping callback types from the recording owner.

* fix(inspector): preserve script identity and parse-failure facts

Fixes #6535

Fixes #6546

* fix(artifacts): recover malformed PYZ sizes without losing siblings

Fixes #6120

* fix(erlang): accept extended BEAM atom lengths

Fixes #6503

* refactor(erlang): reject overflowing BEAM atom lengths

* fix(gpu): preserve PTX and AMD target facts

Fixes #6469

Fixes #6483

* refactor(targets): preserve ambiguous MZ header evidence

* fix(adb): retain refused activity launch reports

Parse complete nonzero Activity Manager reports while preserving other command failures and their captured output. Admit dollar signs only in the activity component name for nested Android classes.

Fixes #6528

Fixes #6527

* fix(mcp): retain partial evidence through shared delivery

* fix(python): honor version-specific code equality

Fixes #6547

* fix(pyinstaller): preserve namespace declarations

Project PYZ namespace packages as metadata so their query paths do not claim output directories. This lets independent CArchive files retain the same logical path and remain inspectable and extractable.

Fixes #6539

* docs: clarify evidence ownership and verification guidance

* docs(adb): clarify emulator observation field

* docs(artifacts): clarify active-reader extraction scope

* fix(erlang): select the last repeated BEAM chunk

Fixes #6518

* fix(il2cpp): distinguish rank-one array types

Fixes #6534

* fix(python): align pyc and PYZ marshal decoding

Fixes #6560
Fixes #6559

* fix(browser): preserve frame identity across captures (#4040)

* fix(browser): preserve capture evidence fidelity

* fix(browser): align frame identity comparisons

* fix(browser): retain historical frame evidence

* fix(electron): retain same-document script owners

* fix(electron): bind frame proof to tree reads

* fix(browser): retain malformed frame scope

* fix(browser): retain WebMCP event scope

* fix(browser): retain malformed timeline identity

* fix(browser): unify frame identity and scope

* fix(electron): prove enclosing frame ancestry

* fix(electron): separate ancestry from file policy

* fix(browser): preserve transient fragment semantics

* fix(browser): retry uncertain initial frame scope

* fix(browser): preserve runtime frame identity

* refactor(browser): separate independent capture fixes

* fix(electron): detect tree-first document replacement

* fix(electron): retain same-document tree authority

---------

Co-authored-by: eco-null <eco-null@users.noreply.github.com>
Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>

* fix(elf): resolve symbol versions from requirements

Fixes #6533

* fix(elf): follow loader version index precedence

VERDEF entries replace VERNEED names for shared indices, matching glibc lookup behavior.

* test(firmware): bind owned producers and exercise real limits

* fix(adb): preserve push target semantics

Fixes #6566

Fixes #6567

* fix(cli): merge sibling output filter slices

Fixes #6542

* fix(browser): preserve Link context semantics

Fixes #6504

Fixes #6505

Fixes #6506

* fix(python): classify historical headers and signed flags

Fixes #6568, #6487

* fix(python): normalize legacy code flags for equality

Preserve CPython 3.7–3.10 constructor flag semantics in code keys.

* fix(environment): compare captured bodies by representation

Keep JSON, text, and retained bytes distinct through fidelity comparison. Compare binary bodies by bytes, including valid UTF-8 bytes against equivalent text, and keep empty-body compatibility explicit.

Fixes #6564

* docs: clarify derived values and comparison controls

* fix(environment): scope fidelity ID normalization

Reserve equal global ID pairs and stop normalizing contested pairings. Keep numeric measurements literal, scope keyed collection identities through learning and comparison, and normalize string references only where learning established the same field as an ID.

Fixes #6565
Fixes #6550
Fixes #6557

* test(process): print the progress frames when the sparse-child bound fails (#6427)

The sparse-child capture test asserted the frame bound with a bare boolean,
so a failure only printed "expected false to be true". The assertion now
carries the running progress details (elapsed time, frames, samples,
interactions) as its failure message. This only adds diagnostics to a
failure; it does not change what passes and does not fix the flake.

Refs #5889.

🤖 Generated with Claude Code (Sonnet 5.5) 🦊

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>

* fix(il2cpp): name the metadata file when linking finds malformed metadata (#6425)

* fix(flutter): honor ELF note alignment when reading build IDs (#6423)

* fix(flutter): honor ELF note alignment when reading build IDs

* test(flutter): key shared note ranges by effective alignment

Cover a program and section declaration that share one note range: differing effective alignments (8 and 4) must both be read, while declarations that both normalize to four bytes (1 and 4) still share one read. 24-byte build IDs pad identically at both alignments, so only the range key decides the read count.

* refactor(firmware): centralize audited tool requirements

* fix(javascript): associate requested bindings only with resolved native addons (#6402)

* fix(javascript): associate requested bindings only with resolved native addons

* test(javascript): cover the exact native addon association

---------

Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>

* fix(python): reject code keys and set members with unhashable consts (#6424)

* fix(python): reject code keys and set members with unhashable consts

* fix(python): refuse to hash a tuple that is still being read

* fix(python): hash a code object's co_consts tuple, not its item list

---------

Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>

* fix(model): recognize named chat templates in summaries

Count retained template content rather than only the default key; keep names-only inventories distinct.

Fixes #6569

* fix(model): retain resolved GGUF alignment under cancellation

Publish alignment from the parser when explicit metadata validates it, or when metadata completion establishes the default. Preserve that fact through later parsing and hashing failures.

Fixes #6579

* fix(environment): reject out-of-range clock offsets

Reject invalid ISO offsets while preserving valid boundary values. Validate stored clock-format offsets through the same parser so malformed imported formats cannot reach rendering.

Fixes #6524

* fix(environment): preserve submillisecond insert ordering

Compare retained submillisecond components after milliseconds, so ordered collections recognize and place distinct timestamps correctly.

Fixes #6525

* fix(environment): restrict epoch normalization to clock fields

Require a time-field token boundary before treating numeric values as epoch clocks, so identifiers such as media format codes remain literal. Clarify recognized field forms and clock offset and precision semantics in the fidelity guide.

Fixes #6551

* fix(process): preserve sibling path identities

Keep aliases on selected paths at clear text delimiters so a longer sibling filename is not rewritten as the executable. Version the normalization contract and cover source and actual PTY behavior.

Fixes #6586

* fix(sigrok): retain recording suffix for autodetection

Copy recordings under a private filename that retains the caller-selected final suffix, and pass that filename to sigrok. This restores filename-based autodetection without exposing the source directory.

Fixes #6580

* test(go): assert unresolved MZ candidate boundary

* fix(pe): ignore unused export ordinal slots

Fixes #6577

* fix(macho): recognize arm64_32 binaries

Fixes #6581

* fix(go): honor amd64p32 pointer width

Fixes #6590

* fix(pe): accept BER indefinite CMS values

Fixes #6574

* fix(browser): reject overlapping indexed source maps

Fixes #6562

* fix(browser): resolve prefixed XHTML form actions

Fixes #6561

* fix(browser): preserve well-known path case

Fixes #6571

* docs(browser): clarify parser evidence contracts

* fix(gpu): accept percent-prefixed PTX entry names

Fixes #6592

* fix(gpu): accept PTX compute target aliases

Preserve the source spelling of NVIDIA's documented compute_xx target synonym.

Fixes #6494

* fix(gpu): parse legacy PTX body parameters

PTX 1.0 through 1.3 declares kernel parameters inside the entry body. Keep later-version parameter-list parsing unchanged.

Fixes #6594

* fix(gpu): keep unknown targets out of exact filters

Fixes #6599

* fix(gpu): resolve extended CUBIN symbol sections

Fixes #6600

* docs(firmware): document the Windows Unblob submodule checkout (#6464)

On Windows, the pinned Unblob commit cannot be checked out because 11
upstream fixture paths under tests/integration/ are invalid on NTFS.
Document the verified sparse-checkout workaround and point Windows
contributors to it from CONTRIBUTING.


Claude-Session: https://claude.ai/code/session_01LjhoPGpqAwHwcCgectzRsp

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* fix(recovery): retain file outcomes and cleanup ownership

Use shared owned-file and directory primitives throughout Wakaru discovery,
snapshots, report parsing and publication. Preserve primary read failures and
completed fingerprints when final close fails or receives cancellation.

Retain exact cleanup capabilities before retiring workspaces, block new work
while cleanup is unresolved, and bind error status to the aggregate owner so
partial cleanup and later retry keep resource attribution current.

* fix(pe): preserve biased export ordinals

Accept the format-derived ordinal range for PE export slots.

* docs: clarify parser authority and compatibility evidence

* fix(python): preserve code-name identity

Fixes #6593

* fix(managed): validate metadata operands and locals declarations

Classify MemberRef operands from parsed signatures, validate signature table references and LocalVarSig blobs, and compare observed locals declaration presence. Keep partial CIL facts and transient-budget ownership at the method reader.

Fixes #6578
Fixes #6572
Fixes #6558
Fixes #6513

* fix(python): preserve historical signed code fields

Keep signed co_stacksize and co_flags for historical magic entries whose constructors accepted them. Preserve magic 3410 with a constructor-revision limitation; reject negative values from 3411 onward.

Fixes #6487

* fix(python): compare code objects without inline caches

Fixes #6597

* fix(tests): stabilize process fixtures and malformed input coverage (#6570)

Preserve renamed-executable identity, complete-line readiness and bounded child lifetimes. Use genuinely malformed Inspector input and explicit ADB file-probe/push/hash facts while preserving cancellation, cleanup, Evidence and completeness assertions. Adopt current main Go fixtures and retain malformed COFF coverage.

Fixes #6556
Refs #6480

* test(apktool): own lifecycle fixtures and process observations

* refactor(mcp): reuse parsed Evidence tool parameters

* fix(runtime): align remaining Node security pins and setup guidance (#6316)

Derive exported Docker image versions from .nvmrc through generated package metadata. Align active setup translations with patched Node release floors and verify actual CLI exports, real Docker/agentenv, package and website workflows.

Eight built-test shards pass; three unrelated managed source failures reproduce on clean main and remain tracked in #6650.

Refs #6480

Co-authored-by: gensokyo-dev <342580529+gensokyo-dev@users.noreply.github.com>

* fix(model): retain exact safetensors shape dimensions

Keep unsigned integer declarations exact until their owning field is interpreted. Bound actual offsets at conversion and validate shape products in declaration order, preserving large empty tensor shapes without hiding an earlier overflow.

Fixes #6627

* fix(python): read legacy TYPE_INT64 marshal values

Fixes #6609

* fix(apk): read constructed BER signer signatures

Share the bounded BER tree reader between PE certificates and APK v1 CMS, then accept constructed OCTET STRING signature values while retaining the declared signer and certificate facts. Embedded X.509 certificate decoding keeps its DER requirement.

Fixes #6603

* fix(adb): count Unicode requested permissions

Count requested-permission rows containing Unicode names, and keep counting subsequent rows in the same dumpsys section.

Fixes #6584

* fix(signals): reject missing sigrok channels

Recognize the exact selected-channel diagnostic even when sigrok-cli exits successfully, while preserving partial producer output in the invalid-input details. Cover both public JSON adapters and add the missing-channel case to the Linux real-tool verifier.

Fixes #6631

* fix(elf): reuse recognized machine architecture facts

Preserve the shared EM_HEXAGON classification in Go and Rust image reports.

Fixes #6629

* fix(asn1): decode four-byte BER lengths unsigned

Accumulate length octets arithmetically so a four-octet length with bit 31 set remains a positive unsigned value and is rejected by the containing-range check rather than becoming a negative length.

* fix(gpu): preserve parameter and target declarations

Decode observed V2 CUDA parameters, retain qualified PTX parameter names and entry resource counts, and preserve family and feature target declarations through selection and consistency checks.

Fixes #6624
Fixes #6625
Fixes #6626
Fixes #6618
Fixes #6613
Fixes #6606

* fix(captures): preserve canonical TShark alias fields

TShark resolves selected field aliases before writing packet JSON. Keep the requested selector in Evidence while accepting the canonical field key produced for it.

Fixes #6595

* fix(qlog): mark records with unknown STREAM FIN incomplete

Keep known STREAM ranges and the unknown FIN value, while listing the source event among incomplete fact records so summary and retained views agree.

Fixes #6598

* fix(pcap): retain absolute HTTP request targets

Reuse parsed URI facts and preserve uncertainty when mixed request records cannot all be resolved from the producer output.

Fixes #6537

* fix(rust): preserve historical cargo-auditable feature facts

Accept observed feature arrays, preserve omitted values as unknown across producer revisions, and keep older retained reports compatible. Reuse bounded JSON sizing before graph expansion.

Fixes #6622

* fix(ptx): normalize parameter declarators once

Fixes #6606

* fix(managed): compare observed call targets

Compare aligned managed call edges by their reported declaring type and method name, keeping missing target facts unknown without comparing build-local tokens. Document the static comparison limits.

Fixes #6630

* fix(artifacts): compare PyInstaller declaration facts

Compare stable producer-tagged declarations independently of materialized payloads. Preserve missing facts as unknown without applying PyInstaller rules to foreign metadata or nested container entries.

Fixes #6484
Fixes #6485

* test(managed): retain derived facts in cloned method fixtures

Clone caller-scoped call and field facts with the fixture methods. Keep known CIL changes alongside unavailable target metadata in the missing-row controls.

* refactor(apk): consume shared BER node facts

Remove the APK reader's pass-through children and content projections so format consumers read the shared BER owner's nodes directly. Keep only the encoded-span helper needed for exact issuer and certificate bytes, and describe CMS BER as this parser's structural owner.

* fix(adb): delimit permission section by indentation

* test(managed): preserve field-access owner identities

* fix(managed): share reference instruction semantics

Use the same parsed method/field operand predicate when producing member edges and comparing call and field shapes. Exclude ldtoken and operands without a value from both interpretations.

* test(managed): align source goldens with metadata validation

Fixes #6650

* fix(native): consolidate observed address validation

* fix(gdb): separate queued work from MI synchronization

Queued console followers have not dispatched MI records. Use active protocol requests when deciding whether the admitted close can verify and detach its inferior.

Fixes #6661

* fix(gdb): wait for the outer console command to finish

After a running result, queue a tokenized read-only MI query to establish that GDB has returned to its command loop. Keep collecting through intermediate inferior stops and retain the query reply as protocol Evidence.

Fixes #6660

* fix(erlang): retain admission cancellation reason

Preserve the caller's AbortSignal reason when BEAM inspection is cancelled before provider admission. Keep the production filesystem boundary and late-completion contracts covered.

Fixes #6658

* fix(desktop): probe GameMaker data filenames

Select conventional Linux and macOS game data files for the existing bounded FORM/GEN8 probe. Add service-boundary regressions and document the selected names.

Fixes #6659

* fix(harmony): exclude raw resources from manifests

Classify known resource paths once and keep a raw resource named module.json from changing the package model or appearing as a manifest. Add a filesystem-boundary regression and clarify the documented hint.

Fixes #6638

* fix(gpu): bound embedded section reads

Fixes #6657

* fix(gpu): retain resources across entry pragmas

Fixes #6651

* fix(reference-source): preserve raw POSIX filename identity

Mark directories with non-UTF-8 names unsupported without following replacement-character paths, while retaining readable sibling observations.

Fixes #6653

* fix(reference-imports): retain exact dynamic import targets

Use the shared exact-literal reader for dynamic imports, preserving expression-free templates and transparent TypeScript wrappers while leaving computed targets unknown.

Fixes #6654

* fix(reference-imports): retain exported import-equals dependencies

Route exported external import-equals declarations through the same relationship owner as plain declarations so their required target is not dropped.

Fixes #6655

* fix(managed): respect native module case semantics

Preserve the observed native format when comparing module identities so ELF case-only differences remain mismatches while unknown platform behavior stays unresolved.

Fixes #6652

* fix(il2cpp): reject duplicate method tokens per image

Reject duplicate MethodDef row identities within one owning image before publishing native addresses or Ghidra seeds, while retaining cross-image reuse and folded code addresses.

Fixes #6610

* refactor(reference-imports): share exact literal parsing

Route legacy nested import-expression handling through the same exact-literal owner as dynamic imports, removing a duplicate string-only check.

Fixes #6654

* fix(model): preserve optional metadata and tensor alignment

Fixes #6656
Fixes #6637
Fixes #6636

* fix(macho): infer legacy Intel simulator platforms

Fixes #6649

* fix(macho)!: preserve ambiguous chained import names

Keep the exact producer expression and report unknown symbol identity when chained-import text cannot distinguish a binding addend from a literal name suffix.

Fixes #6648

BREAKING CHANGE: inspect_macho import names can be null when identity is ambiguous. Use raw_name for the observed expression and name_status to distinguish it from an exact symbol name.

* fix(elf): retain raw export name bytes

Fixes #6647

* refactor(macho): separate dyld import and export parsing

Keep mode-specific inputs and parsed row types at their respective parser boundaries.

* fix(macho): preserve import identity in retained views

Use distinct row contracts for imports and exports, retaining ambiguous names and historical import defaults without adding import-only fields to export pages. Verify the SDK-advertised schemas and both facets through MCP.

* style(desktop): format GameMaker probe coverage

* chore(deps): update Vite security patch and dependency inventory (#6416)

Update Vite/PostCSS and refresh the lock-derived inventory while retaining TSRS platform selection and runtime dependency versions.

Source/static checks, installed-platform signatures, full/production audits, real CLI/MCP workflows and all four package platforms pass. Portable CI retains the independently reproduced main PTX fixture failure tracked in #6675; use the owner-authorized existing PR-only baseline exception. Refs #6480.

* fix(release): isolate npm publication from dependency execution (#6681)

Keep dependency execution in read-only release jobs and validate the original qualified archive with Node built-ins before publication. Refresh registry state on every publisher attempt; retain exact bytes/channel verification, v1 read-only recovery, and installed CLI/MCP checks before MCP publication.

Static/docs checks, artifact/registry/recovery tests and real package/rehearsal verification pass. All four hosted package platforms pass. Portable CI retains the independently reproduced main PTX failure in #6675 and fail-fast missing coverage; use the owner-authorized existing PR-only exception, with limitations documented in the PR. Refs #6480.

* refactor(json): consolidate representation size accounting

Share bounded UTF-8 and UTF-16 JSON measurements across state serving, write retention and source-map admission. Remove duplicate counters and distinguish retained state estimates from encoded bytes while preserving limits and exact refusal diagnostics.

* refactor(android): own search summaries at the producer boundary

Reuse validated class pages for inline pagination audits while retaining complete provider replies. Move retention into the shared Android workflow and remove the MCP-only service wrapper and application text reparser. Preserve unknown JSON metadata, late cancellation and retention failures.

* test: align boundary fixtures with producer contracts

Model GDB console completion, use real directory reads and cleanup for late-enumeration failures, and distinguish supported PTX pointer qualifiers from unknown declarations. Preserve original assertion and cleanup failures through scoped test finalizers.

* fix(rust): retain parsed defaults in cached Evidence views

Borrow sealed reports only when they match the parsed facts; otherwise cache the owned normalized snapshot. Keep historical feature omissions unknown through saved bundles and repeated retained pages without changing the source Evidence.

Fixes #6689

* fix(model): retain parsed headers when hashing is interrupted

Share partial observation ownership across GGUF and Safetensors. Preserve completed header facts through cancellation and MCP error delivery without claiming finalized digests or file identity.

Fixes #6684

* fix(gpu): preserve PTX parameter and target facts

Fixes #6664
Fixes #6675
Fixes #6680
Fixes #6695

* fix(go): read inline build strings from ELF mappings

Fixes #6716

* fix(rust): scan read-only ELF load data

Fixes #6663

* fix(il2cpp): deduplicate consistent ELF registration aliases

Fixes #6715

* fix(il2cpp): track unresolved ELF relocation widths

Fixes #6701

* fix(artifacts): preserve raw directory name identity

Fixes #6665

* fix(desktop): retain unsupported filename coverage

Fixes #6665

* fix: preserve unknown historical module resolution

Fixes #6679

Fixes #6678

Fixes #6612

* fix: retain proven createRequire URL bases

Fixes #6612

Fixes #6678

* fix(apple): preserve keyed archive table identity

Fixes #6662

* fix(plist): reject XML dictionary keys without values

Fixes #6682

* fix(apple): preserve NIBArchive special real values

Fixes #6583

* fix(signals): reject missing targets and invalid UTF-8

Fixes #6671

Fixes #6693

* fix(beam): accept bytes after declared form

Fixes #6717

* fix(artifacts): compare source map target identity

Fixes #6706

* fix(pyinstaller): retain executable member metadata

Fixes #6702

* fix(pyinstaller): decode interned unicode names

Fixes #6676

* fix(pyinstaller): retain colliding archive observations

Keep source-distinct CArchive rows and decoded PYZ children in inventory. Refuse only when extraction reaches a destination conflict.

Fixes #6674

* fix(dotnet): preserve managed metadata comparison facts

Decode VARARG sentinels and allow omitted empty blob heaps while retaining failures for referenced missing data. Compare field and catch targets using owning parsed facts instead of build-local tokens.

Fixes #6720
Fixes #6686
Fixes #6713
Fixes #6712

* fix(dotnet): keep TypeRef catch identity unknown

Fixes #6712

* test(apple): assert projected interface builder facts

* fix(dotnet): correct metadata comparison regressions

* fix(il2cpp): deduplicate aliased registration fields

Fixes #6715

* test(pyinstaller): match interned name producer encoding

Use Python 3.13 PYZ magic with the matching archive version and exercise FLAG_REF with TYPE_INTERNED.

Fixes #6676

* test(dotnet): split managed comparison blocks

* test(apple): isolate NIBArchive real projection group

* fix(parsers): correct Rust import and PTX alignment control

* fix: recognize aliased createRequire bindings

Fixes #6612

Fixes #6678

* test(apple): retain NIBArchive limitation path

* fix(il2cpp): deduplicate aliased metadata registration

Fixes #6715

* test(managed): retain observed target changes across member remaps

* fix: track require.resolve references

Fixes #6678

* test(pyinstaller): encode PYZ list length correctly

Write the marshal list length as its raw little-endian count and surface inventory limitations when completeness regresses.

Fixes #6676

* fix(managed): retain unknown catch identity beside observed name changes

* fix: retain require.resolve source facts

Fixes #6678

* refactor(javascript): name retained require call facts explicitly

* fix: resolve selected require call members

Fixes #6678

* fix(javascript): keep unresolved created-loader bases unknown

* refactor(artifacts): centralize completed file outcomes

Project completed values and cancellation from the descriptor owner once across artifact consumers. Preserve completed void outcomes and format-specific validation and cleanup policies.

Keep later aborts from replacing captured APK and SQLite close failures, and retain the exact APK cancellation reason. Exercise real descriptors, public workflows and cleanup retries.

* fix(javascript): isolate package facts from host prototypes

Build frozen package field snapshots with a null prototype and restrict reference-source main lookup to own properties. Preserve declared null and empty values, index fallback, source identity and unavailable manifest evidence.

Compare inherited main and exports lookup with the native Node resolver, retain package observation coverage and verify CLI/MCP package workflows.

* fix(captures): advertise offline capture inspection without a browser provider (#6725)

* fix(parsers): preserve format boundaries and query occurrence identity

Normalize parsed facts at their owning boundaries across PE exports and certificates, ELF section names, GPU bundles, IL2CPP generic types, dyld slice selection, source maps, environment replay, and JavaScript mapped arguments. Preserve partial facts and explicit unsupported cases; clarify the corresponding documentation.

Validation: 831 affected tests passed with one existing skip across 23 files; build, typecheck/lint, documentation, formatting, and 2133-file test discovery passed. Format controls are synthetic; real Unity, GPU, and ARM-device execution is unverified.

Fixes #6709
Fixes #6688
Fixes #6719
Fixes #6718
Fixes #6704
Fixes #6703
Fixes #6692
Fixes #6723
Fixes #6687
Fixes #6726
Fixes #6696
Fixes #6691
Fixes #6727

* fix(process): require observed environment ownership tokens

Centralize run-token extraction at the environment observation boundary and require an own field. Build decoded OS environment records without host prototypes while retaining source names, empty values and last-assignment precedence.

Reject inherited matching tokens before cleanup signaling, preserve normal owned-group behavior, and verify live Linux process environments alongside native CLI/MCP capture workflows.

* fix(parsers): retain independent facts and boundary identity

Share bounded ELF section-name decoding across native and note readers; retain read-only ranges and coverage reasons when names are unknown. Preserve ELF symbol-version bytes, dyld-bound method uncertainty, partial gRPC-Web trailers, and per-parent model identity checks. Compare GPU signatures as bytes and require a PTX module header. Share selected-path errno classification across EVM and BEAM.

Validation: 397 shared-parser tests passed (one existing skip), 5 MCP schema tests passed, 14 EVM error-mapping tests passed, and 23 protobuf CLI/MCP matrix cases passed across focused runs. Build, typecheck/lint, docs, formatting, and 2134-file discovery passed. Full EVM provider workflow remains locally unverified on Darwin arm64; format controls are synthetic.

Fixes #6738
Fixes #6737
Fixes #6745
Fixes #6740
Fixes #6749
Fixes #6747
Fixes #6739

* fix(setup): preserve source keys across configuration boundaries

* fix(setup): keep the CLI-only skill choice offered under a client ID (#6401)

* fix(setup): keep the CLI-only skill choice offered under a client ID

* test(setup): cover the accept-all CLI-only skill selection

* fix(apple): report a read-denied selected archive as host access denial (#6479)

* fix(apple): report a read-denied selected archive as host access denial

inspect-keyed-archive reported a host read refusal on the caller-selected bundle entry as artifact_operation_failed/execution_failure with reason io, leaving EACCES or EPERM only inside message text. The caller selected that entry, so the refusal is a selection outcome with its own typed classification, system code and read-access recovery.

DirectoryArtifactReader now marks a refused entry open with DirectoryArtifactEntryAccessDenied, which retains the host system code and the selected entry path. KeyedArchiveReader, which owns the caller's selection, maps it to the shared AnalysisAccessDeniedError, and artifact failure translation keeps that typed error instead of replacing it with a generic io failure. Other readers and other reader failures are unchanged; every existing reason keeps its current projection.

Adds the access_denied projection coverage plus a real host refusal through the production reader and a non-permission control.

* fix(apple): carry cleanup ownership and cover raw refusals for a denied selection

Addresses the two review findings on the denied-selection classification:

- The typed host access denial now carries the reader's cleanup observation, so a denied open that left a resource owned does not project as fully cleaned.
- Adds coverage that starts from raw host EACCES/EPERM/EIO refusals with the real DirectoryArtifactReader open path, so the marker itself is exercised rather than constructed by the test. EIO stays an artifact failure.

* fix(native): report plist symlink loops and overlong paths as selection failures (#6371)

Co-authored-by: Abelo9996 <Abelo9996@users.noreply.github.com>

* docs(cli): document opt-in Bash completion setup (#6250)

* fix(doctor): check only the current platform's Hopper launcher candidates (#6253)

* fix(doctor): check only the current platform's Hopper launcher candidates

Doctor built one launcher list for every platform, so on macOS it could
report a Linux-only path as a healthy Hopper ahead of a renamed bundle,
and setup then registered that path. Give per-platform automatic
candidates one owner that runtime configuration and doctor both read,
scan renamed macOS bundles and probe the Homebrew cask only on macOS,
and share runtime's renamed-bundle scan with the doctor host.

Fixes #6228

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(doctor): keep the renamed Hopper bundle scan asynchronous

Sharing runtime's synchronous scan made the doctor host read
~/Applications with readdirSync, which blocks the event loop on a slow
or network-mounted home. Give MacHopperPaths an asynchronous scan for
doctor beside runtime's lazy synchronous one, sharing the directory
filter and candidate builder, and check that both return the same
candidates.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(doctor): model renamed launchers with runtime's POSIX joins

Since #5235, runtime builds macOS launcher paths with posix.join, so on
Windows the discovery table's native-join renamed launcher no longer
matched runtime's candidate. Model launchers with posix.join and keep
filesystem setup native, as MacHopperPaths.test.ts does.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

* fix(entropy): name offset when the byte window starts past EOF (#6495)

Fixes #6482

Co-authored-by: Abelo9996 <Abelo9996@users.noreply.github.com>

* fix(sqlite): report result-column capacity refusals as capacity (#6327)

A projection wider than SQLite's native result-column limit failed with the
generic "too many columns in result set" SQL error. The inspector could not
attribute that error to a generated column, so the worker reported "format"
and the parent blamed the caller's path as invalid_format even though the
selected database was readable.

Classify the refusal from SQLite's own column limit for the private
connection, carry the selected and engine counts through the worker reply,
and project it as a capacity resource constraint. The database stays
readable, no partial rows are returned, and the recovery names the column
counts and splitting the projection across calls.

* fix(web): refuse an existing script-export destination as invalid input (#6313)

* fix(web): refuse an existing script-export destination as invalid input

export_web_scripts refused an existing output directory or link correctly,
but reported it as artifact_operation_failed / execution_failure with no
field issue or recovery. SafeOutputTree.create raises that refusal as a
path failure before any output tree exists, and the export mapper wrapped
it as an artifact failure.

Attribute pre-admission path refusals to output_directory as an
AnalysisInputError, as environment export already does for its directory.
Both now share one helper owned beside SafeOutputTree; failures after the
tree is admitted keep their artifact semantics.

Fixes #6029

* fix(artifacts): type SafeOutputTree destination refusals

Admission conversion treated any cleanup-free path failure raised before
admission as a refusal of the selected directory, so an acquired-tree setup
failure, such as being unable to restrict a created root to 0700, blamed
valid output_directory or directory input.

SafeOutputTree.create now raises SafeOutputTreeDestinationRefusal for the
refusals of the selection itself: a relative or reserved root, an existing
root, and an accepted existing root that is not an empty directory. The shared
conversion accepts only that type, directly or wrapped by
SafeOutputTreeCreationFailure, so script and environment export no longer
track admission separately.

* fix(evm): reject repeated container keys in Solidity metadata (#6230)

* fix(evm): reject repeated container keys in Solidity metadata

* test(evm): check the repeat-key diagnostic for deeply nested keys

* docs(ci): add a repeatable static analysis example (#6496)

* docs(ci): add a repeatable static analysis example

* fix(ci): isolate example reports and repair the guide link

---------

Co-authored-by: gongfpp <22683417+gongfpp@users.noreply.github.com>

* fix(setup): retain unresolved environment facts when deriving budgets

* fix(parsers): preserve format evidence and caller input identity

Retain bounded FAT and offset-tagged PE observations through archive hashing,
classification, and content nodes. Preserve PyInstaller member paths according
to the observed carrier dialect and share APK v1 signing-name semantics.

Read SQLite's implicit sqlite_schema table without inventing a stored schema
declaration. Distinguish unsupported ELF numbering and decoded-name admission
from malformed data and fixed scan allowances. Preserve Sigrok basenames and
classify matching explicit input-format refusals at their owning boundary.

Validation: 554 distinct tests passed, 3 skipped across serial focused runs;
build, typecheck/lint, docs, formatting, and discovery (2141 files) passed.
Sigrok command-boundary simulators do not establish real-tool decoding.
Small repeated public PyInstaller CLI runs had equivalent inventory facts;
warm median elapsed 0.5431s before / 0.5426s after, with no speed claim.

Fixes #6733
Fixes #6741
Fixes #6722
Fixes #6732
Fixes #6736
Fixes #6734
Fixes #6764
Fixes #6766
Refs #6739

* fix(parsers): respect container facts and real decoding boundaries

Use zip.js's parsed symlink fact for inventory and open refusal. Treat an
unmatched Godot trailer as an unconfirmed candidate while preserving failures
for confirmed malformed packs. Report entropy path-resolution failures at the
selected-input boundary.

Separate Go header discovery from inline string extents in ELF and Mach-O,
and apply the aggregate content allowance after validating real source ranges.
Normalize binary plist UID widths through the existing scalar-copy owner,
budget exact clone growth, preserve source identity, and refuse inexact UIDs.

Validation: 389 distinct tests passed, one Linux /proc test skipped on Darwin;
build, typecheck/lint, docs, single-thread formatting, and discovery (2142 files)
passed. CLI/MCP Go fixtures and public ZIP/Godot/keyed-archive workflows ran.
Real producer prevalence and external provider execution were not established.

Repeated ordinary keyed-archive CLI inventory facts and estimated reply tokens
were equal before/after. Warm median elapsed was 0.5166s / 0.5177s; no speed claim.

Fixes #6751
Fixes #6763
Fixes #6754
Fixes #6767
Fixes #6762
Fixes #6779

* refactor(mcp): consolidate Evidence recording and operation progress

* fix(setup): reuse URL component admission for Hopper releases

* perf(cli): avoid loading analysis commands for version reports

---------

Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>
Co-authored-by: Hifza Zafar <152292315+hifzabuildsai@users.noreply.github.com>
Co-authored-by: Can <159528514+eco-null@users.noreply.github.com>
Co-authored-by: eco-null <eco-null@users.noreply.github.com>
Co-authored-by: Cassie_Fox <249835913+CassieFox23@users.noreply.github.com>
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-authored-by: Wade <160517347+Sriram-PR@users.noreply.github.com>
Co-authored-by: RRFZ <247457381+Plastic-time@users.noreply.github.com>
Co-authored-by: pablodiazjorge <102741971+pablodiazjorge@users.noreply.github.com>
Co-authored-by: Ezra Brilliant <64943463+ezrabrilliant@users.noreply.github.com>
Co-authored-by: gensokyo-dev <342580529+gensokyo-dev@users.noreply.github.com>
Co-authored-by: GamerDubz <105694977+GamerDubz@users.noreply.github.com>
Co-authored-by: raidnouredden-ux <raidnouredden@gmail.com>
Co-authored-by: 吴佳浩Alben <35291372+XiaoBinGan@users.noreply.github.com>
Co-authored-by: Abel Yagubyan <abelyagubyan@berkeley.edu>
Co-authored-by: Abelo9996 <Abelo9996@users.noreply.github.com>
Co-authored-by: ShuaixinHuang <1836656514@qq.com>
Co-authored-by: Paritosh Lahre <155938586+SandstromPL@users.noreply.github.com>
Co-authored-by: Dery Ferdika <114888155+DeryFerd@users.noreply.github.com>
Co-authored-by: gongfpp <1360187968@qq.com>
Co-authored-by: gongfpp <22683417+gongfpp@users.noreply.github.com>
2026-10-24 00:06:03 +08:00
Ezra BrilliantandClaude Opus 5.5 6d6cbdf8c4 fix(doctor): report a missing --format value as MISSING_OUTPUT_VALUE (#6489)
`mcp doctor --format` and `--format=` reported VALIDATION_ERROR
"Invalid output format: missing", although docs/cli.md documents
MISSING_OUTPUT_VALUE for valued output flags without a value and
`capabilities` already returns it. Classify absent and empty values as
missing, share the existing missing-value message, and keep
VALIDATION_ERROR for a present but unsupported format.

Fixes #6455


Claude-Session: https://claude.ai/code/session_01LjhoPGpqAwHwcCgectzRsp

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 14:39:41 +08:00
Wadeandgensokyo-dev d11fca46fc fix(cli): print compact mcp doctor JSON when stdout is redirected (#5514)
rea mcp doctor rendered json with Formatter.format, which always indents, so
redirected --json, --full-output and validation errors stayed indented. The
dispatcher now passes the stdout TTY state and the doctor renders json and
jsonl through jsonOutputParts like the other commands.

Fixes #5435

Co-authored-by: gensokyo-dev <342580529+gensokyo-dev@users.noreply.github.com>
2026-10-23 02:23:39 +08:00
584937c761 feat(mcp): make MCP startup phase budgets configurable (#6304)
* feat(mcp): make MCP startup phase budgets configurable

REA_MCP_INITIALIZE_BUDGET_MS, REA_MCP_TOOLS_LIST_BUDGET_MS,
REA_MCP_FIRST_CATALOG_BUDGET_MS, and REA_MCP_DOCTOR_DEADLINE_MS keep
today's defaults and refuse values outside 1 to 2147483647 ms. Codex's
30 second startup figure stays client guidance.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(mcp): enforce the first-catalog budget while phases are pending

The packaged startup probe applied each phase timeout on its own and
checked REA_MCP_FIRST_CATALOG_BUDGET_MS only after both requests
finished. A shorter first-catalog budget now clamps the in-flight
request and cancels it from one deadline measured at startup.

Co-authored-by: morluto <morluto@users.noreply.github.com>

---------

Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: morluto <morluto@users.noreply.github.com>
2026-10-22 23:54:01 +08:00
morluto 2589c0074d fix(mcp): attribute doctor timeouts to owned deadline facts 2026-10-22 22:13:31 +08:00
bdf0be87c6 fix(mcp): name response-byte and startup ceilings in recovery (#6005)
* fix(mcp): name response-byte and startup ceilings in recovery

Name REA_MCP_MAX_RESPONSE_BYTES, its effective value, the 10485760 byte
floor, and process restart when the stdio response budget is hit, and
state that the setting does not raise Node's string ceiling. Name fixed
MCP startup phase budgets when those deadlines fire, and keep Codex's
30 second startup figure as client guidance.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(mcp): distinguish string ceilings and keep dossier recovery

Name REA's reserved string limit separately from Node's ceiling, restore
detail:summary discovery for an oversized complete evidence bundle, and
apply the doctor deadline sentence when binary_session itself times out.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(mcp): keep retained SDK errors inside the text budget

Quoted SDK validation diagnostics expand when recovery JSON is escaped
into MCP text. Shorten only that preview so the retained-evidence
reference still arrives as parseable error content.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(mcp): keep export advice when a summary exceeds the response budget

An oversized get_evidence_bundle summary no longer repeats detail: summary.
export_evidence_bundle still writes the retained bundle outside one MCP frame,
including when a larger byte budget cannot raise the string ceiling.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* style(mcp): format the SDK diagnostic preview return

oxfmt rejected the fitted-preview ternary in EvidenceMcpServer.ts, which failed Static checks and the CI required gate.

Co-authored-by: morluto <morluto@users.noreply.github.com>

---------

Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: morluto <morluto@users.noreply.github.com>
2026-10-22 05:48:56 +08:00
morluto c5fb86d194 fix(mcp): bind catalog identity to output schema profiles
Include the selected output advertisement profile while retaining canonical output validation in the identity. Reuse the SDK output projection and select doctor expectations from both profiles. Correct the compact input error documentation.
2026-10-15 01:41:31 +08:00
187f113f79 fix(identity)!: preserve invocation uncertainty and advertised catalogs
Capture the current entrypoint once without inventing a path. Preserve unknown registration ownership through setup and doctor, refuse unsafe registration plans, and retain exact pinned-launcher readiness checks. Hash the same full or compact schema projection advertised by each MCP server and use that identity in doctor comparisons.

BREAKING CHANGE: server.command_path may be null when the invocation path is unavailable. Registration and setup statuses, and doctor classifications, may report unknown.

Reviewed-PR: #3230, #3015
Co-authored-by: qorvr <330186599+qorvr@users.noreply.github.com>
Co-authored-by: Hemang Varshney <218273197+hemv-857@users.noreply.github.com>
2026-10-14 16:10:46 +08:00
morlutoand1052326311 ae433f0983 fix(mcp): bound retained doctor stderr while draining output
Retain a 64 KiB diagnostic prefix while counting and draining the complete stderr stream. Report truncation and avoid manufacturing a replacement character at a truncated UTF-8 boundary.

Validation: all nine doctor tests pass, including real noisy child, malformed bytes, canonical catalog and startup deadline controls.

Based on #3121.

Co-authored-by: 1052326311 <65798732+1052326311@users.noreply.github.com>
2026-10-14 13:58:17 +08:00
morluto 80de3145e4 fix(mcp): align advertised contracts and retain workflow evidence
Use one named input-schema projection for clients and catalog identity, preserve dialect declarations, and bound compact schema depth. Keep canonical SDK output validation when output advertisements are disabled. Release request-owned state on cancellation and carry upstream limitations through composed workflows. Preserve doctor full-output envelopes and validate navigation facts at their owner.

Reviewed fixes: #2088, #2108, #2513, #2684, #2693, #2695, #2836, #2864, #2868, #2877, #2933, #2998, #3002, #3015.

Validation: focused schema, SDK output-validation, context, and workflow tests passed; no schema dialect or canonical output validation was relaxed.
2026-10-14 08:32:00 +08:00
morluto 7fc0b5ec9c fix(mcp): join owned doctor shutdown without signaling saved PIDs 2026-10-13 15:12:08 +08:00
Ysopwnable ea92328775 fix(mcp): validate doctor probe against its discovered output contract (#1461)
Validate the target-free probe against the named definition discovered from the complete production catalog. Preserve initialized server and inventory observations when the SDK rejects the probe, and retain the schema failure reason, absolute deadline and owned-child cleanup.

The negative stdio replay regression fails on original main and passes the fix; actual production-boundary fault injection confirms rejection. Nine doctor/help tests, serial full static checks and real small-target CLI/MCP parity pass. Successful doctor reports remain byte-identical. No performance or model-token reduction is claimed.

Closes #1454. Refs #1065, #723, #1433.
2026-10-10 14:33:57 +08:00
morluto b7d803dd10 fix(cli): drive mcp doctor help from a single option schema
Parses rea mcp doctor arguments against one zod option schema that is also the help renderer's source of truth, answering --help/-h in any position without starting a diagnostic session and keeping actionable errors for unknown options and invalid formats. Updates the landed help assertion to the schema-rendered format list and adds boundary coverage for help positioning and --format edge cases.
2026-10-09 08:00:07 +08:00
morluto a69b8337ab fix(cli): support MCP doctor help 2026-10-09 07:29:16 +08:00
morluto ecf3ed8bdb refactor(tools): remove remaining arbitrary workflow caps 2026-09-29 17:00:06 +08:00
morluto 0d96275b68 refactor: return complete tool results inline 2026-09-29 07:08:46 +08:00
morluto b8a9da0c0f refactor(mcp): make analysis states unrepresentable (#468)
* refactor(config): parse capability settings into exact states

* refactor(runtime): encode replay and process outcomes exactly

* refactor: make analysis states unrepresentable

* refactor(mcp): register canonical schemas with SDK

* fix: preserve capability and catalog invariants

* fix(build): remove stale compiled modules

* fix(replay): update real verifier policy dependency

* fix(package): verify the complete MCP catalog

* fix(test): supply complete replay reload policy

* refactor(contracts): centralize output schema lookup

* refactor(mcp): delegate protocol boundaries to SDK

* fix(browser): approve real source capture

* fix(mcp): preserve catalog guidance
2026-08-10 02:28:11 +08:00
morluto fecddc0549 perf(mcp): reduce cold-start work 2026-07-26 15:02:58 +00:00
morluto 035e1144b0 feat(server): advertise available tools dynamically 2026-07-24 13:18:58 +00:00
N0zoM1z0 d25e98c3f9 feat: complete REA remediation program 2026-07-22 22:09:12 +09:00