Group repeated session tool availability and provider capability facts, preserve complete named contracts and ordered limitations, and remove duplicate catalog descriptors. Update consumers and migration guidance.
Validated by all 31 CI jobs, focused source checks, and actual Claude Code/Linux Hopper/native Windows Ghidra small-target workflows on the current integration tree.
Refs #723.
BREAKING CHANGE: binary_session no longer includes server_identity.catalog.tools; tool_availability and provider capabilities use grouped facts documented in docs/mcp-contracts.md.
Co-authored-by: Kaoru0822 <208281322+kaoru0822-kitauji@users.noreply.github.com>
Replay the complete real Apple dylib-use output through the native provider boundary and retain rejection of an empty repeated options field. Align skill-install Result assertions and CDP initiator prerequisites with current contracts. The producing parser fix is already retained from main.
Refs #1970
Co-authored-by: factory-droid[bot] <138933559+factory-droid[bot]@users.noreply.github.com>
Co-authored-by: Kaoru0822 <208281322+kaoru0822-kitauji@users.noreply.github.com>
Give atomic Evidence and snapshot writes an explicit resource owner, preserve primary failures and completed publication receipts, and retry retained staging cleanup at the owning lifecycle. Verify inode identity before publishing or deleting; unknown path disappearance remains a cleanup failure. Keep snapshot admission reusable when a session closes and reopens.
Consolidate JavaScript source-scope classification so cancellation cannot fabricate failures for unselected files. Classify configuration failures as invalid input. Remove stale assertions and reuse shared test cleanup ownership.
Validation: 266 focused tests before final ownership tightening, 108 affected tests afterward, static and documentation checks, actual invalid-configuration CLI behavior, and isolated test temporary-directory hygiene.
Refs #1968
Expose the selected Mach-O slice and its container offset, derive that identity once, and map entrypoint and nonempty segment Evidence to container coordinates. Reject relocation sections whose byte length is not divisible by the entry size instead of silently accepting a floored count.
Validation: native parser and selected-architecture boundaries, actual public CLI and stdio MCP on a compiled universal Mach-O, and a pyelftools 0.33 malformed-table probe. Full Linux pwntools bridge verification was unavailable on this host.
Refs #1973
Parse Apktool YAML and XML at the producer boundary, preserving scalar spelling, XML text and partial string coverage. Retain repeated JADX inventory rows with explicit duplicate-name counts and lookup ambiguity. Make the Apktool verifier report cleanup failures.
Validation: parser, provider and JADX fixture boundary tests; typecheck, lint and documentation checks. Real Apktool and JADX engines were not exercised.
Refs #1978
Interpret device paths once as POSIX paths and apply only the selected host's filename constraints. Preserve original PM paths and diagnostics, reject duplicate destinations before pulls, and use exact-name filesystem reservations to detect case and Unicode aliases without guessing volume rules.
Consolidate failed-package cleanup ownership and strengthen no-output assertions. Validated 44 provider fixture tests, 47 filename/parser tests, six compiled CLI tests, static/docs checks, and package checks. No Android hardware was used.
Remove repeated MCP input parsing after the SDK has validated named contracts. Reject unmodeled fixture operations and locate prompt contracts by name so unrelated catalog ordering cannot select a contract.
Validated focused browser, Electron, runtime-observation and MCP workflow tests, advertised schemas, packaged MCP and static checks. Application-boundary validation remains in place.
Reject malformed UTF-8 ZIP filenames at the decoding boundary while retaining CP437, Unicode Path overrides and leading U+FEFF. Preserve valid CustomAttribute parent and value facts when constructor identity is invalid, with an explicit partial-coverage issue.
Validated ZIP integrity and filesystem-provider workflows, managed producer cases, temporary-file hygiene, package and static checks. Removed the superseded producer-only invalid-constructor regression.
Thread the existing resource scope through skill and client configuration reads. Keep failed-close handles owned, wait for admitted reads, and surface unresolved cleanup alongside doctor, uninstall and update results.
Validated focused setup/doctor/update and filesystem boundary checks, temporary-file hygiene, packaged setup/update/uninstall workflows, and static checks. No external provider installation was performed.
Separate observed relationship coverage from general provider caveats. Authenticate and replay coverage without changing legacy Evidence IDs; expose unknown and capped relationship sets through CLI/MCP graphs and traces.
Validated focused domain, snapshot, IDA/Ghidra boundary, MCP, schema, package and static checks. Provider fixtures establish contract behavior; licensed engine behavior was not reverified.
* fix(wasm): preserve debug module names in WABT output
* fix(native): accept otool dylib-use options
Preserve modern dylib options while requiring one encoding metadata row and retaining duplicate and multiline-output checks. Share field parsing between preservation and validation.
Verify delayed dylibs through real macOS CLI/MCP workflows and refresh the high-address golden with the provider's -h -l capture. Validated 85 focused native tests, npm run check, and verify:macos-bundle on macOS arm64.
---------
Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>
* fix(wasm): accept unpadded WABT section rows
WABT 1.0.42 right-aligns section names in a nine-column field. DataCount
fills that field, so its objdump row has no leading whitespace. Accept
that legal row. Header recognition, the required field shape, and section
range checks still reject malformed output.
Fixes#1956
* fix(ci): refresh MIPS expectations and Windows test paths
Assert admitted soft-float ABI metadata and its committed storage policy through the public target resolver, retain FP64 rejection, and point the curated Windows lane at the moved filesystem reconstruction tests.
Validation: reproduced all four prior failures; 97 focused tests and check:ci pass. Native Windows and real Ghidra execution were not performed locally.
---------
Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>
Remove 30 redundant truthy-JSON schema checks. Validate advertised groups and property parity across every primary command, retain required-field checks, and assert structured malformed-input errors. Exercise JEB positional validation rather than an unknown flag.
Validation: 13 public CLI adapter tests, combined checks, and temporary-directory hygiene. Host configuration refusal coverage from main is preserved.
Validate Mach command counts, byte totals, scalar uniqueness, section counts, and build-tool counts before claiming exhaustive evidence. Preserve raw output in typed parsing failures and distinguish dylib commands from the dynamic loader. Correct inconsistent handcrafted fixtures and label their provenance.
Validation: 38 native tests, combined checks, and temporary-directory hygiene. Multiline dylib-name ambiguity was reproduced with the host Apple otool; regression workflows use captured output and injected command runners.
Require observed request method and initiator metadata instead of inventing defaults. Preserve valid redirect history while excluding later events for malformed reused IDs. Keep unavailable or invalid encoded lengths unknown and retain valid zero values.
Validation: 57 CDP boundary tests plus temporary-directory hygiene and combined checks. Protocol events come from the fake CDP server; no live browser verification is claimed.
Rebind display paths only for matching artifact identities in primitive and composed replay. Preserve semantic Evidence IDs, payloads, and historical snapshot provenance. Classify snapshot filesystem workflows as boundary tests.
Validation: cache, direct workflow, MCP replay, and Evidence tests; combined checks and temporary-directory hygiene. Provider interactions are simulated for replay verification.
Wait for directory operations, close owned streams and traversal handles, and retain the mounted reader when cleanup fails so close can retry before detach. Consolidate stable-artifact filesystem regressions into the boundary suite.
Validation: 35 artifact tests and temporary-directory hygiene, including real hdiutil diagnostics. Directory-close and traversal races use injected faults.
Track attempted writes, preserve original content and modes from one regular-file observation, and continue rollback after individual failures. Return tagged failure details and retained backup paths through setup.
Validation: setup filesystem and recovery tests, combined compile/typecheck/lint/format/metadata/architecture checks. Filesystem fault injection covers backup, ambiguous publication, and rollback failures.
The ADB provider, which the apktool provider mirrors, had the same two
defects: a relative REA_ADB_PATH escaped as UNKNOWN because selection
ran before the try that maps configuration failures, and a missing adb
binary projected generic target advice instead of its reason.
Move the configuration-failure-to-refusal mapping into one
providerConfigurationRefusal helper used by both providers, resolve the
ADB selection inside the mapped try, and share the CLI refusal assertion.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>
Give snapshot capture, header admission, packed members, and container streams one file-handle owner. Serialize adoption with reader shutdown, retain failed closes for retry, and remove snapshots only after descriptor cleanup succeeds.
Cover unread members, pending opens, zero-byte members, and paired format/cleanup failures. Keep intentionally unrecoverable test fixtures hygienic and strengthen managed identity and process cleanup-authority expectations.
Validation: full npm run check; 99 targeted tests and temporary-directory hygiene; exact discovery of 877 test files. No real vendor-provider verification was performed.
* feat(ghidra): admit soft-float MIPS32r2 ELF with integer-register float storage
Soft-float o32 executables (Linux kernel builds, bare-metal firmware) were
refused because their ABI flags record declares no CPR1 and FP ABI 3.
Ghidra's MIPS32 compiler specifications model only hard-float o32, so typed
float/double signatures (DWARF) would be bound to $f12/$f14/$f0 and the
decompiler loses the real dataflow: `return sel ? a : b` decompiled to
`return in_f0;`, and callers passed phantom FPU inputs.
Admit ELF32 ET_EXEC o32 MIPS32r2 targets whose ABI record declares soft-float
with no CPR1, and commit `mips_soft_float_storage: o32-integer-registers-v1`
in their analysis profile only (hard-float profile digests are unchanged).
ReaGhidraMipsSoftFloat.java runs after auto-analysis and, for each function
with a float/double formal type, either keeps the types with the storage
Ghidra's o32 model assigns to a same-size integer (pair alignment, stack
spill), or, for auto-parameters such as hidden struct-return pointers, swaps
float types for same-size integers under dynamic storage. No function is left
on hard-float storage. The report is written last; the provider client
requires it exactly when the profile commits the policy and refuses the
session before any operation otherwise.
Verification: real Ghidra 12.1.4 lane (verify:ghidra:mips) now builds
hard- and soft-float fixtures in both byte orders and checks exact
parameter/return storage through inspect_native_api, caller constants,
snapshot profile and Evidence outcome; test:fast, verify:package and an
installed-tarball soft-float run pass.
* fix(ghidra): model soft-float MIPS as a prototype for every typed callable
Review feedback on #1949: rewriting storage per Function left external
functions and indirect calls through typed function pointers on Ghidra's
hard-float o32 prototype, which still read $f12/$f14/$f0 while the session
reported success. A negative control confirms it: without the change,
`(*rea_mips_soft_callback)(8.5f, 9.5f, sel)` decompiles to
`(*rea_mips_soft_callback)(in_f12, in_f14, sel)`.
ReaGhidraMipsSoftFloat.java now derives a `__rea_soft_float` prototype from
the program's own default o32 model (encoded, float pentries and
float-consuming rules removed, so byte-order-specific aggregate rules and the
hidden struct-return pointer are kept), installs it as a specification
extension, and assigns it with dynamic storage to every non-thunk function,
external function and function-pointer definition whose formal types use
float or double. Thunks follow their targets. Float types are kept
everywhere, so the integer-retype fallback is removed. The script fails
without writing its report if the derived prototype or any reassigned
function still uses an FPU register.
The policy becomes `o32-soft-float-prototype-v1` with a
{functions, definitions} report. The soft-float fixture adds an indirect
call through a typed pointer; the real-Ghidra lane requires its constants
with no FPU inputs and the struct return's `float scale` in $a1.
* test(ghidra): pin ODDSPREG acceptance for soft-float MIPS ABI records
Review on #1949 suggested requiring flags1 == 0 for soft-float, since MIPS_AFL_FLAGS1_ODDSPREG names CPR1 registers. GCC 5.2 kernel modules, LLVM/Rust firmware and Clang -msoft-float objects all record flags1 = 1 with FP ABI soft-float: it is the MIPS32 o32 toolchain default and selects nothing without CPR1. Keep accepting it, say why at the check, and test that soft-float with ODDSPREG is admitted while any other general flag is refused.
Replace duplicate path-helper checks with real ZIP inventory coverage and replace global memory sampling with bounded-chunk assertions while retaining 64 MiB digest coverage. Await artifact client cleanup, preserve Hopper lease ownership during teardown, and expose typed application failures.
Capture exact filesystem identity and nanosecond timestamps, validate descriptor/path state before and after streaming, and enforce the captured byte count. Bound growing files to their captured size plus one byte using precise positional reads; keep unrepresentable display sizes explicitly unknown.
Reuse the captured-state comparison around Apple assetutil and retain existing reader cleanup ownership. Validation: focused reader and affected consumer suites, full static/architecture checks, and a compiled directory-read probe on Bun 1.3.14 passed. Includes same-size mutation, growth, truncation, cancellation, FIFO replacement, cleanup retry, ZIP identity forgery, ASAR and Mach-O coverage.
Serialize resolution with shutdown, retain the owned browser after close failures, and retry cleanup before another launch. Wire cleanup through the shared service and CLI/MCP lifecycles, preserving primary failures and completed native reports alongside cleanup diagnostics.
Separate the browser resolution workflow from lifecycle ownership. Validation: seven resolver source tests, eleven application tests, MCP shutdown retry, public CLI parity, and full static/architecture checks passed. Real Chromium behavior remains unverified.
Move artifact inventory, application projection, runtime reconciliation, and TODO scanner suites into their real boundary lanes. Extract Hopper process/socket cleanup cases while keeping injected session cleanup cases in composition.
Validation: all 56 focused tests passed; discovery finds all 876 test files exactly once. Test assertions and provider verification scope are unchanged.
Move the six JavaScript application suites that read production artifacts or launch compiled analysis workers out of composition. Focused test planning now prepares their runtime prerequisites instead of treating them as source-only tests.
Validation: all six relocated suites and adjacent artifact/reference suites pass (70 tests across eight files); discovery finds every test exactly once.
Bracket symlink target observation with the identity captured during enumeration. Report replacements and disappearing links as changed, and keep cancellation and I/O mapping at one owner boundary.
Remove duplicate target-error wrappers and unreachable external-target handling. Validation: reference reader filesystem suite passes, including deterministic replacement during observation.
Catch binary selection failures at the adapter boundary and share the domain error projection with apktool. Missing and relative adb paths now report capability_unavailable with host configuration guidance in both message and remediation.
Validation: ADB/apktool provider suites and public CLI refusal suites pass. No connected-device behavior changed or verified. Refs #1952.
Supply schema-validated inspect_analysis_view arguments after completed native function Evidence is recorded and exceeds MCP delivery capacity. Preserve the full canonical result and its provenance. Verify exact source paging, export and following-call usability through real Ghidra stdio MCP.
Fixes#1849
Co-authored-by: Minghang Chamling <133689405+cold-atom@users.noreply.github.com>
Co-authored-by: factory-droid[bot] <138933559+factory-droid[bot]@users.noreply.github.com>
Co-authored-by: Kaoru0822 <208281322+kaoru0822-kitauji@users.noreply.github.com>
A missing apktool launcher projected the generic "unavailable for the
current target" message and remediation, leaving the actual reason only
in details.reason. A relative REA_APKTOOL_COMMAND escaped as UNKNOWN
because command selection ran before the try that maps configuration
failures.
Resolve the selection inside that try for both execution and
availability, and pass the refusal reason as userMessage, as JADX and
the adapters fixed in #1224 do.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* test(ida): expect callee edges from the call graph on non-x86 targets
* fix(ida): read callee edges from the call graph on non-x86 targets
* test(ida): pin address order of call graph additions
* docs(ida): describe the call graph source for non-x86 callees
* test(ida): pin the x86 baseline, the call graph reader and its failure modes
* fix(ida): check the call graph root, scope truncation to the root and degrade without callgraph
* test(ida): expect managed PE off the x86 path, possible-truncation wording and a complete ARM64 traversal
* fix(ida): keep managed PE and x86 reads off the call graph path and qualify truncation
* test(ida): answer only expected leaf roots and assert the requested call graph roots
* test(ida): expect a call graph the producer cannot give consistently to fail the callee read
* fix(ida): fail the callee read when the call graph cannot be obtained consistently
* fix(ida): preserve call edge authority and workflow uncertainty
* fix(ida): preserve explicit root call graph caps
---------
Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>
Analyze exact selected sources in first-request order before unrelated
static scopes, preserving full artifact identity and byte verification.
Retain selection in CLI/MCP Evidence and summaries, with explicit unknowns
and partial coverage for unselected sources. Keep default full analysis.
Fixes#1848
Co-authored-by: factory-droid[bot] <138933559+factory-droid[bot]@users.noreply.github.com>
Co-authored-by: Kaoru0822 <208281322+kaoru0822-kitauji@users.noreply.github.com>
Verify retained provider operation health before reusing the same target and profile. Retire failed Hopper clients through existing owned cleanup, preserve healthy reuse and cancellation, and start replacement children lazily on the next query.
Refs #1719. Actual headless Linux Hopper child-death/reopen recovery and 85 related tests passed against the current-main merge tree; Ghidra/IDA liveness and the historical snapshot report remain outside this verification.
Co-authored-by: sx4im <sx4im@users.noreply.github.com>
Co-authored-by: N0zoM1z0 <161784452+N0zoM1z0@users.noreply.github.com>
Co-authored-by: Kaoru0822 <208281322+kaoru0822-kitauji@users.noreply.github.com>
* fix(install): describe automatic setup and guard truncated downloads
The help text said the installer only installs the CLI and that
`rea setup` must be run separately, but the installer starts
`rea setup` interactively whenever a terminal is available. Describe
that behaviour and the --no-setup/--no-prompt opt-out.
Wrap the installer body in `main` and call it on the last line, so a
truncated `curl | bash` download fails to parse instead of executing a
partial script. Behaviour is otherwise unchanged.
Add an installer scenario that runs a copy truncated right after the
npm install line and requires that it prints nothing and never invokes
npm; it fails against the previous script.
Validation: bash -n, shellcheck 0.10.0, installScript boundary tests.
* fix(install): require the closing token before running main
Wrapping the installer in main() kept a truncated download from running
a partial body, but the last line was a bare `main "$@"`. If the piped
download stopped right after `main` (or `main `), bash saw a complete
command at EOF and ran main with no positional parameters. Under the
documented `curl ... | bash -s -- --dry-run` (or --version / --no-setup)
path that silently dropped the caller's flags, so a dry run could resolve
the latest release, run npm, and start setup.
The call now sits in a brace group:
{
main "$@"
exit $?
}
Bash does not run the group until it has read the closing brace, so
every truncation inside it is a syntax error and nothing executes; a
truncation before the group leaves only function definitions. A comment
in install.sh records why the bare call is not enough.
Validation: a new boundary test pipes install.sh to `bash -s --
--dry-run` truncated at every byte offset from the end of main() to EOF
and asserts that npm and the rea binary are never invoked; offsets
inside the group must exit non-zero with a bash parse error and no
stdout, offsets before it must print nothing, and the complete group
must finish the dry run. Reproduced the original bug by cutting the old
script after `main`: the dry run invoked npm install. `bash -n` and
shellcheck 0.11.0 are clean.
* fix(install): wrap the whole script in the truncation guard
The previous guard only wrapped the final `main "$@"` invocation, so the
preamble streamed on the `curl ... | bash` path before any incomplete
compound command was open. A download cut off right after the bytes
`set` reached EOF with a complete, argument-less `set`, which prints every
shell variable, so inherited tokens could reach stdout.
The brace group now opens on line 2, directly after the shebang, and
closes on the last line, enclosing `set -euo pipefail`, the constants, the
functions and the final `main "$@"` / `exit $?`. Bash must therefore read
the whole file before running anything; any earlier cut is a syntax error
at EOF. `set -euo pipefail` still applies inside the group (checked
manually for `false`, an unbound variable, and a failing function body),
and shellcheck 0.11.0 and `bash -n` are clean.
The truncation test now pipes every prefix of install.sh, from byte 0 to
EOF (about 6,100 cuts, run 16 at a time with a minimal environment that
includes a sentinel variable), into `bash -s -- --dry-run`. Before the
group opens it must exit quietly; inside the group it must fail with a
syntax error; in every case stdout stays empty and neither npm nor rea is
invoked. Only the complete script performs the dry run.
* fix: order remaining outputs independently of host locale
String.prototype.localeCompare without a pinned collation follows the
host's ICU default locale, so the same input could be ordered differently
on different machines. 614fb768 moved canonical domain projections to
compareUnicodeCodePoints; provider, application, catalog, and setup
orderings still used locale collation.
Order every remaining source site by Unicode code point, including Mach-O
symbols, firmware extraction paths, ASAR entries (previously pinned to
"en" collation), evidence ID sets, browser and Electron target, script,
and resource inventories, reconstruction obligations, capability and
provider listings, and setup environment details.
The current tool and prompt catalog names order identically under both
comparators, so the catalog identity digest is unchanged.
* fix(artifacts): order ASAR entries by their portable path
@electron/asar lists host-spelled paths, so code-point order over the raw
spelling still differed by host: "a\b" sorts after "a0" on Windows while
"a/b" sorts before it on POSIX. Pair each listing with its portable
archive path once, order by that path, and keep the host spelling for
@electron/asar calls and adapter keys.
MCP annotations derived destructiveHint only from mayDiscardData, so tools
that run the analysis target or a caller-selected executable
(capture_process_scenario, observe_native_calls) and tools that change an
attached Android device (install/uninstall/push/start/stop) advertised
destructiveHint: false.
Add explicit executesForeignCode and mutatesDevice effect flags, audit them
across all 170 tool effect records, and derive
destructiveHint = mayDiscardData || executesForeignCode || mutatesDevice.
readOnlyHint semantics are unchanged.
Hopper/Ghidra metadata writers (set_address_name, set_addresses_names,
set_bookmark, set_comment, set_inline_comment, annotate_native_function)
can replace or clear existing analyst names and comments, so they now
declare mayDiscardData as well.
A contract test pins the destructive set and requires every tool whose
description mentions running, launching, executing, starting, installing or
uninstalling to declare one of the new flags or carry a reasoned exception,
so wording and effects cannot silently drift. The binary_session effects
output schema carries the new flags.
Focused and zero-match search_android_classes responses embedded every
examined class name: JadxSession records each raw list_classes page into
observed calls, those calls ride into Evidence raw_result on the
registration path, and the complete inline Evidence is serialized into
every tool response (issue #1850: 382,435 bytes for one match).
When the server session retains Evidence, record the complete raw-call
record first, then deliver a summary Evidence that keeps the pagination
audit (totals, offsets, page sizes, elided item counts), the unchanged
normalized matches and truthful coverage metadata, and links the
retained complete inventory by Evidence ID. Sessions without retention
keep the complete inline delivery, so complete CLI JSON output is
unchanged and no provenance is discarded.
🤖 Generated with AI assistance (disclosure voluntary per repo policy)
