* feat(#1517): support custom reviewer instances for /gsd:review
Add a bounded review.reviewer_instances config surface so one model-capable
adapter (e.g. opencode) can run as several independent reviewer identities in a
single /gsd:review pass. Instances participate only via review.default_reviewers,
expand before built-in slugs, are available iff their cli is detected, and a
non-matching entry is a hard error (typo must be loud). >=2 same-cli instances
emit a shared-adapter caveat in REVIEWS.md. Default path with no instances is
byte-for-byte unchanged.
Single-source instance->cli resolution lives in resolveReviewerSelection /
normalizeReviewerInstances (parity-locked in
tests/review-reviewer-instances.test.cjs). cli validated against
KNOWN_REVIEWER_SLUGS only (never arbitrary shell); model/agent opaque, never
shell-interpolated.
Closes#1517
* chore(#1517): backfill changeset pr:1766
---------
Co-authored-by: review-bot <review-bot@gsd>
The claude-global install now provisions .gsd-source (#1477), whose content
is the install-time absolute path to the package commands/gsd source tree.
That path is the checkout/CI workspace root, not the temp HOME, so it is never
normalized to <HOME> and its hash varies by environment. Exclude it from the
parity manifest as a volatile metadata file, same rationale as
gsd-install-state.json.
Address re-review minors on #1487:
- bin/install.js: replace the silent .gsd-source marker catch with a
console.warn so a failed write is diagnosable (walk-up also fails on the
Claude-global layout, so a swallowed error still breaks /gsd-surface).
- tests/runtime-artifact-layout.test.cjs: add a writer fault-injection case
(mock fs.writeFileSync to throw for the marker) proving the catch branch is
reachable, install stays non-fatal, and the warning is emitted.
ADR-1235 step 1: route the trivial-converter runtime group (cursor, windsurf, augment, trae, codebuddy) off the inline install() agent loop onto the descriptor-driven installRuntimeArtifacts path. Establishes the converter-context foundation (pre-converter cross-cutting + no agent-stamp). Agent install output is byte-identical for all 16 runtimes (golden-parity, global + verified local). cline deliberately excluded (local rules-only). Closes#1763.
* fix(#1693): don't double-quote $CLAUDE_PROJECT_DIR-anchored hook paths on Windows
The installer's #2979 legacy-node rewrite ran every managed node hook path
through JSON.stringify on Windows. For local installs the path already carries
a "$CLAUDE_PROJECT_DIR"-anchored quoted prefix, so stringifying produced
"\"$CLAUDE_PROJECT_DIR\"/...". Node then received an argument starting with a
literal " , treated it as relative, and failed MODULE_NOT_FOUND — breaking
every node managed hook at once (a self-locking PreToolUse-guard deadlock).
projectLegacySettingsHookCommand now emits an already-anchored token verbatim
and only JSON.stringify-quotes bare absolute paths (which may contain spaces).
Scoped to win32 so non-Windows token-shape behavior is unchanged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* chore(#1693): add changeset
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
ADR-1239 Phase B (parent #1679). Replace copyWithPathReplacement's 13
hardcoded `const isX = runtime === 'x'` flags + two ~100-line per-runtime
if/else converter chains with a module-level RUNTIME_CONTENT_DISPATCH table
(one entry per runtime: md transform, mdSkipGenericRewrite, mdReattributeAfter,
mdTomlRenameOnCommand, js transform) + a uniform dispatch loop that applies the
cross-cutting steps (path rewrite -> attribution -> stamp -> normalize) once.
Byte-identical install output for all 16 runtimes (golden-parity harness #1730);
codex-verified the transform order + every per-runtime quirk (gemini .toml
rename, copilot/antigravity skip-generic + reattribute, qwen/hermes inline
swaps, .cjs/.js fall-through) is preserved.
Also folds in a pre-existing bookkeeping fix (no-defer): gsd-core/bin/lib/
cli-skew-check.cjs (tsc build artifact from #1755) was eslint-ignored but
missing from .gitignore — added for consistency with the other built artifacts.
Closes#1758
Co-authored-by: review-bot <review-bot@gsd>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
ADR-1239 Phase B (parent #1679). getDirName was a hand-maintained 15-branch
if-chain mapping each runtime to its local content-rewrite dot-dir. Relocate
those values into a documented runtime.localConfigDir descriptor field; derive
getDirName from registry.runtimes[id].runtime.localConfigDir (fallback .claude).
- 16 capability.json gain runtime.localConfigDir (byte-identical values)
- capability-validator.cjs requires it (non-empty dot-dir); registry regenerated
- docs/reference/capability-manifest.md documents the field + the three
divergent values (copilot=.github, antigravity=.agents, kimi=.kimi-code)
- drift-guard test: golden value map + key-set equality both ways
Byte-identical install output for all 16 runtimes (golden-parity harness #1730).
Closes#1756
Co-authored-by: review-bot <review-bot@gsd>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* feat(#1754): CLI version-skew detection — warn when a global install shadows project-local GSD
Addresses #1754 (approved-enhancement). Detects when the running gsd-tools.cjs
is outside the project root while a project-local install exists — the shadowing
scenario from #1748 where a stale global canary CLI (retired @gsd-build/sdk)
silently overrides project-local GSD.
Implementation (Node CLI entry-point, not shell snippet — avoids bloating 93
workflow files past their size caps):
- src/cli-skew-check.cts: pure function checkCliSkew({resolvedPath, projectRoot,
projectLocalExists}) → string|null. Compares paths via path.relative; returns
a warning when the resolved CLI is outside the project root AND a project-local
install exists. Includes @gsd-build/sdk removal hint when the path matches.
No I/O (pure), no gsd-sdk literal (avoids bug-2801 lint).
- gsd-core/bin/gsd-tools.cjs: wired at startup via the existing findProjectRoot
resolver. Non-blocking (try/catch; advisory stderr warning, never gates).
- eslint.config.mjs: registers the new ADR-457 generated artifact in the ignores.
- tests: 6-case suite (skew/no-skew/legacy/normalization); all green.
- Golden fixtures regenerated (UPDATE_GOLDEN=1) for the new compiled artifact.
- docs/how-to/update-gsd.md: Diátaxis reference note for the skew warning.
Full suite: 3354 pass, 0 regressions (1 pre-existing local AGENTS.md failure).
lint:ci green.
Closes#1754
* chore(#1754): backfill changeset pr placeholder
* chore(#1754): regenerate INVENTORY-MANIFEST for the new cli-skew-check source module
---------
Co-authored-by: review-bot <review-bot@gsd>
* docs(#956): address adr-phase-coverage findings on the MemPalace capability proposal
Resolves the findings from the adr-phase-coverage audit (matrix + interpreted
gaps posted on #956) by embedding durable decision→phase ownership and the
cross-doc gating item into the proposal, so no deliverable sits ownerless
between phases:
- §15.1 Decision → Phase ownership: every §10 decision + user-facing capability
is the explicit responsibility of one phase. Cross-cutting policies get a
primary owner (D6 onError:skip → P1 manifest-encoded; D3 transport → P2
MCP-primary rendering, P5 CLI-fallback headless).
- §15.2 Dependencies & gating items: Phase 6 is GATED on ADR-857's Migrate
phase (workflows calling loop render-hooks) — recorded as a traced gating
item, not prose. UX-auto + UX-curator have no other wiring surface; if
ADR-857's Migrate is descoped, Phase 6 is formally blocked, not silently
dropped. De-risk: Phases 1–5 ship the full manual-invocation value ahead.
- Phase 3 gate: explicitly verifies the inherited UX-enable surface
(gsd capability enable mempalace + config-set), not just the internal
state resolver.
- Phase 6 gate: names the user-observable automatic surface (a
/gsd-execute-phase run auto-produces MEMORY-RECALL.md at plan:pre, curator
spawns at ship:post) instead of the wiring mechanism.
- §17 open questions: each is traced to the phase whose acceptance must
resolve it (wing-identity→P0, replace-migration→P4, curator-tier→P2,
headless-MCP→P5, phase-6-dep→§15.2 gate, diary-namespacing→P6).
Docs-only (proposal refinement); no runtime change.
Closes#956
* docs(#956): correct the Phase-6 framing — ADR-857 is released, auto-fire is wired and verified
Retracts the 'Phase 6 GATED on ADR-857 Migrate' framing introduced in the prior
commit. ADR-857 (capability system + loop render-hooks + workflow call sites) is
released; the host-loop workflows call loop render-hooks at each canonical point,
so mempalace auto-fires when mempalace.enabled. Verified end-to-end: 'gsd-tools
loop render-hooks plan:pre --raw' with mempalace.enabled:true returns the
mempalace-recall step (capId: mempalace, produces MEMORY-RECALL.md).
- Phase 6 row: 'Loop wiring (shipped via ADR-857)' with the verified gate.
- §15.1 Phase 6 row: wired via shipped ADR-857 infra (no gate).
- §15.2: rewritten from 'Dependencies & gating items' (false premise) to
'Loop wiring status' — documents the released/shipped state + the retraction.
- §17.5: 'Phase-6 dependency' → 'Loop wiring (resolved — shipped)'.
The §15.1 decision→phase ownership matrix, Phase 3 UX-enable gate, and §17
open-question traceability from the prior commit stand (those were accurate).
---------
Co-authored-by: review-bot <review-bot@gsd>
Post-merge coverage-audit follow-ups to epic #1702 (found by an independent
gpt-5.5/high audit + adr-phase-coverage cross-reference). None are CRITICAL —
the 9-rule enforcement shipped and works; these close completeness/integrity
gaps between ADR-1703's promises and the as-built reality.
1. drift-guard bin/install.js scope (ADR-1703 L114-119): the drift guard
covered src/runtime-homes.cts only; the ADR named bin/install.js too. Phase
6's glob expansion made bin/install.js a covered surface. Extended
tests/portability-vocab-drift.test.cjs with TWO sound checks: (a) any
bin/install.js top-level function that directly returns path.*() must be in
PATH_RETURNING_FNS (tight, 0 FP — the body-contains heuristic is unsound
here, ~33 FPs); (b) a curated two-way existence lock on the installer path
helpers (catches a rename making a vocab entry stale; keeps the curation in
sync with PATH_RETURNING_FNS). The residual new-resolver boundary (temp-var
shape) is documented.
2. ci-test-scope wiring (the Phase 6 portability selection rule was
ineffective): eslint-rules/ was not in the product-code prefix list, so an
eslint-rules-only change set code_changed=false and CLEARED the matched
tests (reproduced: targeted_tests=[]). Added eslint-rules/ to the prefix
list and the P1-P4 RuleTester suites to the selection rule (it previously
listed only P5/P6). Verified: code_changed=true, 11 tests selected.
3. ADR-1703 acceptance note amended to record the two further as-built
divergences: the disable-ban shipped as an out-of-band test (not the
specified local/no-portability-disable meta-rule — the test runs outside
ESLint so it cannot be self-disabled, at least as strong); and the
drift-guard bin/install.js scope resolution above.
Epic #1702 all eight phase boxes now checked. No runtime change; no-changelog
(contributor tooling + docs).
Closes#1749
Co-authored-by: review-bot <review-bot@gsd>
ADR-1703 Phase 7 / epic #1702 closeout. The mechanical teardown is already
complete across phases 1-6 (regex scanner retired P3, ratchet deleted P4,
allowlist portability-usage gone, windows-portability-ok comments swept,
every DEFECT.WINDOWS-* predicate rewritten per-phase). This phase delivers the
two remaining ADR-mandated closeout items:
- docs/contributing/adding-a-portability-rule.md: the forward architecture
recipe (Diátaxis Explanation) — the five seams (rule / portability-vocab /
platform-guard / disable-ban / ci-test-scope), the zero-escape-hatch contract,
the shipped-rule catalog, and the step-by-step checklist for adding a new
local/* portability rule.
- docs/adr/1703: Status Proposed -> Accepted (all seven phases shipped); a
Phase 7 as-built note recording the two deviations from the Phase 0 catalog —
Phase 6's rename-only scope decision (#1740) and the codex-review precision
tightening on require-fs-op-fallback.
- docs/contributing/cross-platform-portability-rules.md: cross-link to the new
guide from the rule reference.
No code, no test, no runtime change. Epic #1702 Phase 6 box checked; Phase 7
box + epic closure to follow at PR merge.
Closes#1744
Co-authored-by: review-bot <review-bot@gsd>
* feat(#1740): require-fs-op-fallback production AST rule + Windows transient-lock retry (Phase 6)
ADR-1703 Phase 6 of the cross-platform portability epic (#1702). Adds the
second production-code portability AST rule + the ADR-mandated glob expansion
to bin/install.js and scripts/build-hooks.js.
- eslint-rules/require-fs-op-fallback.cjs: flags an unguarded fs.rename /
fs.renameSync (the atomic-publish primitive named first in
DEFECT.WINDOWS-FS-OPS.symptom) that is NOT inside a try/catch whose handler
references a transient errno ('EPERM'/'EBUSY'/'EACCES' or a *RETRY_ERRNOS
set) AND NOT behind a Windows platform guard. A catch that silently swallows
or cleans-up-and-rethrows without an errno check does NOT satisfy the
defect's 'never silently swallow' clause. copyFile/unlink are deliberately
not flagged (they are the fallback primitives per the defect's own
fix-forward). Scope narrowed to rename per Phase 5's precision discipline;
documented on #1740.
- src/shell-command-projection.cts: export retryRenameSync(from, to) — the
drop-in bounded-retry helper over the existing atomicRenameWithRetry.
- 27 bare fs.renameSync sites across 11 modules routed through retryRenameSync
(capability-lifecycle/lock/source, installer-migrations, milestone, phase,
planning-workspace, roadmap-upgrade, runtime-hooks-surface, state,
workstream). Idempotent on POSIX; resilient to AV/indexer transient locks
on Windows.
- eslint.config.mjs: register rule at error on src/**/*.cts; new focused
portability-rules block covering bin/install.js + scripts/build-hooks.js
(ADR-1703 L124-126 glob expansion — both files are compliant: zero
rename violations).
- tests: 15-case RuleTester suite; portability-rule-disable-ban extended
(PROTECTED_RULES + scans bin/install.js/build-hooks.js with shebang
handling); ci-test-scope portability-lint selection rule.
- CONTEXT.md DEFECT.WINDOWS-FS-OPS predicate rewritten to point at the rule;
docs/contributing/cross-platform-portability-rules.md reference + how-to.
Closes#1740
* chore(#1740): backfill changeset pr:1742
* fix(#1740): tighten require-fs-op-fallback precision (codex review HIGH-1/HIGH-2)
Addresses two false-negative findings from the codex (gpt-5.5/high)
adversarial review of PR #1742:
HIGH-1 — a catch that REFERENCES a transient errno but only rethrows (no
retry/fallback) was marked compliant. The DEFECT.WINDOWS-FS-OPS fix-forward
requires retry, not just recognition. Fix: catchHandlerHasRetrySignal now
requires a loop `continue` backedge OR a `return <call>` delegation; a bare
rethrow is flagged. The misleading `/* retry logic */` valid test is replaced
with a real retry loop, and the rethrow-only shape is added as invalid.
HIGH-2 — the nested-try ancestor walk treated an OUTER errno-catch as
protecting the rename even when an INNER catch intercepted/swallowed the error
(the outer catch is unreachable). Fix: isInsideTransientErrnoTryCatch now stops
at the NEAREST enclosing TryStatement WITH A CATCH HANDLER whose block contains
the rename (try-finally is skipped — it doesn't catch); outer catches are no
longer consulted. The unsound nested-try valid test is converted to invalid,
and a try-finally-skipped valid case is added.
Verified: 17 RuleTester cases pass; zero new production violations (the 27
fixed sites use retryRenameSync; the real retry loops — atomicRenameWithRetry,
capability-ledger/consent, build-hooks — remain compliant via continue/errno);
lint:ci green; disable-ban + vocab-drift green.
---------
Co-authored-by: review-bot <review-bot@gsd>
* feat(#1733): normalize-path-in-content production AST rule (Phase 5)
ADR-1703 Phase 5 — the first production-code rule. local/normalize-path-in-content
(src/**/*.cts, @typescript-eslint/parser): flags a path-returning fn result
(path.basename excluded — returns a separator-less filename) interpolated into an
@-reference / config-dir markdown body without .replace(/\\/g,'/') normalization,
per RULESET.CONTENT-PATH-NORMALIZATION / DEFECT.WINDOWS-PATH-LEAK-IN-MARKDOWN-CONTENT.
Build-and-assess found the canonical defect site (computePathPrefix) already
compliant and only 1 src/ hit — a false positive (path.basename in a status
message) — eliminated by narrowing (exclude basename; require a real @-ref/
config-dir marker, not bare .md). 0 src/ violations: clean forward-prevention.
The out-of-band disable-ban now scans src/**/*.cts too (typescript-estree) so the
production rule also cannot be eslint-disabled. Registered (error) + PROTECTED_RULES;
CONTEXT.md predicates + how-to doc updated.
- RuleTester suite (26 cases)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(#1733): add changeset for Windows agent-skills path-leak fix
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix: harden mutation-matrix.cjs stdin read against EAGAIN on non-blocking pipe
scripts/mutation-matrix.cjs read piped stdin via readFileSync(process.stdin.fd).
On macOS libuv marks the stdin pipe fd non-blocking, so a synchronous read can
throw EAGAIN before the writer fills the pipe — intermittently, under heavy CI
shard load — aborting the script (status 2) and flaking mutation-matrix-ratchet.
Replace with readStdinSync(): an fs.readSync loop that retries on EAGAIN (1ms
synchronous Atomics.wait yield), stops on 0-byte/EOF, and rethrows other errors.
Deterministic regression test injects EAGAIN via an fs.readSync monkeypatch.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* ci: re-run golden-install-parity on src/lib + installer changes (close drift guard)
golden-install-parity hashes every installed bin/lib/*.cjs per runtime, so it
must re-run whenever the built lib could change. ci-test-scope selected it for
neither src/** nor installer changes, so a source-only edit (e.g. #1691's
milestone.cts/roadmap.cts) recompiled bin/lib and silently drifted the golden
fixtures past the scoped lane. Add golden-install-parity.test.cjs to both the
'TS runtime sources' and 'installer and package layout' selection rules, with
behavioral regression tests for each.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: review-bot <review-bot@gsd>
* refactor(#1734): extract install engine from bin/install.js (ADR-1239 Phase B deep move)
Relocate the runtime-artifact install cluster out of the 12,490-line
bin/install.js into a dedicated src/install-engine.cts -> install-engine.cjs:
installRuntimeArtifacts, uninstallRuntimeArtifacts, installOpencodeFamilySkills,
and their cluster helpers (_copyStaged, snapshot/restore, legacy migration,
GSD-entry pruning, preserve/restoreUserArtifacts, OpenCode-family converters,
USER_OWNED_ARTIFACTS).
- bin/install.js imports the engine and re-exports the moved symbols for
back-compat; getCommitAttribution STAYS in install.js (impure config I/O +
argv explicitConfigDir global) and is injected via a resolveAttribution param.
- 17 test files migrated to import the moved symbols from the engine.
- Bookkeeping: eslint built-artifact ignore, .gitignore, INVENTORY manifest+row,
CONTEXT.md Install Engine Module glossary seam.
Behaviour-preserving: install output is byte-identical for all 16 runtimes
(golden-parity harness #1730) — the only delta is the new install-engine.cjs
file shipping in the installed gsd-core/bin/lib/ tree.
Closes#1734
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(#1734): backfill changeset PR number (#1735)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: review-bot <review-bot@gsd>
* fix(#1580): exclude 0/999 sentinels from milestone-complete guard and roadmap analyze
Closed#1445 added the `^999` backlog-sentinel exclusion to the progress
denominators but missed two other resolvers, leaving two user-facing failures
live on a milestone whose only directory-less ROADMAP heading is a backlog
sentinel:
(A) `milestone complete` was blocked by the unstarted-phase guard in
src/milestone.cts — it flagged `### Phase 999: Backlog` as an unstarted
phase and refused to close a fully-shipped milestone without --force.
(B) `roadmap analyze` (src/roadmap.cts) counted the sentinel in phase_count
and routed `next_phase` straight into Phase 999.
Both now skip Phase 0 (pre-milestone) and Phase 999 (backlog) sentinels,
mirroring the engine-wide convention (phase-id getMilestoneFromPhaseId,
roadmap-command-router SENTINELS, the #1445 progress filters). Symptom (C)
(state.cts total_phases) was already fixed inline by #1445/#1514 and is out
of scope here.
Regression coverage folded into tests/fix-1445-*.test.cjs (scenarios C and D);
updated tests/bug-978 fixture to use a real unstarted phase (Phase 2) instead
of a 999 sentinel, since the sentinel is now correctly excluded from that guard.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* chore(#1580): add changeset for 0/999 sentinel exclusion fix
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Promotes the verify-time enforcement decision that accreted as four chronological addenda on ADR-550 (#1259/#1279/#1346/#1278) into a single first-class architecture-of-record. ADR-550 keeps the spec-phase contract; this ADR owns the test-tier enforcement producer (locate -> machine-prove-fail-first -> run -> dispose) in src/prohibition-enforcement.cts.
Verified against live src/*.cts: the green AND-gate lives in runProhibitionEnforcement, the empty-file vacuity guard is isNonVacuousNodeTestPass (red-side is isNonVacuousNodeTestRed), and the ADR-857 pointer cites the real 'Verification substrate vs. plug-in tier' section / decision #6.
Closes#1606.
Appends an 'Alternatives considered (recall / representation / packaging side)' addendum to ADR-550: the two NEW spec-phase-side rejections (the withdrawn LLM requirement classifier #652, and a deterministic prohibition-recall engine) plus cross-references to the two already-decided ones (no polarity field on truths — D3; deferred single dispatcher CLI — D7e). Enforcement-side alternatives stay in ADR-1606.
Closes#1607.
Adds the greppable PROBE.*/PROHIB.* single-line predicates ADR-550's Consequences promised alongside the CONTEXT.md glossary. RESCOPED from the original issue: the glossary entries (Probe Family / Probe Core / Edge Probe / Prohibition Probe / Verification Tier / Verification substrate) ALREADY landed (CONTEXT.md), so only the predicate block was net-new — the draft's premise that the glossary 'never landed' came from a grep that hit 'binary file matches' on a multibyte char. Research-derived N17/N18 numbers deliberately kept out of machine-canon (they live hedged in docs/design/verifier-reach.md).
Closes#1608.
ADR-1239 Phase B (parent #1679). Safety net for the upcoming engine deep-move:
captures the COMPLETE emitted install output of all 16 runtimes as a golden
baseline so the move PR can prove byte-identical parity.
tests/golden-install-parity.test.cjs spawns the real installer per runtime into
a temp HOME, normalizes the temp path to <HOME>, excludes the two volatile
metadata files (gsd-file-manifest.json, gsd-install-state.json — the only
run-to-run variance after normalization, empirically), SHA-256s every remaining
file, and asserts the manifest matches tests/fixtures/golden-install-parity/<rt>.json
(8,957 hashes total). UPDATE_GOLDEN=1 regenerates; mismatches list
added/removed/changed paths. Non-vacuous (corrupting a hash fails the run).
Closes#1730
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* refactor(#1727): derive NON_CLAUDE_RUNTIMES from the capability registry
ADR-1239 Phase B (parent #1679). NON_CLAUDE_RUNTIMES was a hand-maintained
15-element literal whose own doc-comment said "keep in sync with bin/install.js
and getDirName()" — a parallel source of truth that can drift from the
capability registry. Derive it instead:
Object.keys(capabilityRegistry.runtimes).filter(id => id !== 'claude').sort()
The exported value is byte-identical to the old literal (the registry's
runtimes key set minus claude is exactly the 15 entries), so there is no
observable behavior change; the list can no longer drift from the registry.
capability-registry.cjs is a committed, dependency-free data module (no cycle).
Drift-guard test: golden-oracle deepEqual (non-circular) + a role-based
cross-check from the registry metadata + every member must have a non-'.claude'
getDirName branch (a registry runtime missing a getDirName branch now fails CI).
Closes#1727
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(#1727): backfill changeset PR number (#1728)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(#1727): put docs-exempt marker on its own line so parse.cjs extracts it
DOCS_EXEMPT_RE is line-anchored (^...$ + m flag); the marker only counts on
its own line. It was appended to the end of the body text, so it was never
extracted and docs-lint failed in CI (fail_docs_missing). Verified via direct
parse.cjs extraction (docsExempt now non-empty, marker stripped from body).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* feat(#1724): complete install write-confinement (copyWithPathReplacement, installCodexConfig)
ADR-1239 Phase B (parent #1679). PR #1706 (2a) confined the layout-driven
plan path and _copyStaged's inline guard; this completes the destSubpath
write-confinement acceptance criterion for the two remaining write sites
and canonicalizes _copyStaged.
- copyWithPathReplacement: new required confinementRoot param; a fail-closed
gate (assertDestWithinConfigHome + hasExistingSymlinkBetween) runs BEFORE
the rmSync/mkdirSync; root threaded through recursion + all 4 call sites
(stageRoot for pristine staging, targetDir for the 3 install sites); writes
go through the validated absolute path. Exported for behavioral testing.
- installCodexConfig: confines config.toml, agents/, and per-agent
agents/<name>.toml (name from agent frontmatter) via the canonical gate +
symlink-escape guard (parity with the other two functions).
- _copyStaged: fail-closed when configDir omitted (all callers pass it);
delegates strict-subpath to the canonical gate, keeps its symlink guard,
writes through the validated absolute path.
Reuses the existing assertDestWithinConfigHome (handles absolute dests via
path.resolve) and hasExistingSymlinkBetween — no new module. Behavioral
regression tests (escape/dest==root/fail-closed/symlink/name-injection),
red-first proven; cross-platform symlink tests use t.skip not bare return.
Closes#1724
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(#1724): backfill changeset PR number (#1725)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* feat(#1708): typed documentation-sourced #853 dispatch-flatten
Graduate the #853 orchestrator-backgrounding decision from a scattered RUNTIME==='codex' prose check to a typed, documentation-sourced engine decision. Adds a backgroundDispatch dispatch sub-axis (sourced per host: codex+cursor documented true, 9 documented false, 5 undocumented), shouldFlattenDispatch(dispatch) (inline UNLESS background && backgroundDispatch, fail-closed), and a gsd_run query dispatch-should-flatten the plan/execute workflows call. Cursor is newly background-eligible per its docs (inline->background) — a documentation-justified behavior change. No RUNTIME-name residue for this decision.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(#1708): backgroundDispatch citations in matrix + CONTEXT note
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(#1708): address review findings on typed dispatch-flatten
Code/adversarial review: convert the manager.md/autonomous.md Compound Action preamble from hardcoded 'On Codex' to FLATTEN-based branching (the handlers already use the query; the preamble contradicted them and was wrong for cursor); make shouldFlattenDispatch null-safe + type-honest (accepts raw 'undocumented' registry values); make backgroundDispatch a required descriptor field (matching its siblings, all 16 carry it); strengthen the config.runtime behavioral test; update the bug-853 prose-pin test + comment. Security review clean; Codex confirmed no fail-open.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(#1708): backfill backgroundDispatch in role:runtime test fixtures
Making backgroundDispatch a required descriptor field broke role:runtime fixtures in capability-manifest-version/capability-registry/host-integration-descriptors tests that build a dispatch object without it (caught by full gsd-test, not scoped npm test). Backfill backgroundDispatch:false into the well-formed fixtures; the deliberately-malformed 'required-field' test fixture is left malformed by design.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(#1708): update fix-1521 dispatch-gating assertion to the FLATTEN gate
fix-1521 pinned the codex-specific run_in_background prose that #1708 graduated to the typed dispatch-should-flatten/FLATTEN gate. Update its assertions to verify FLATTEN=false gating (not a runtime name) + that the old RUNTIME===codex gate is gone. Caught by full gsd-test.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(#1708): add changeset for typed dispatch-flatten
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(#1708): remove stray temp PR-body file
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(#1708): add issue ref to bug-853 allow-test-rule annotations
ADR-456 requires every allow-test-rule exemption to carry a see #NNN reference; the source-text-is-the-product annotations added when migrating the prose assertions lacked it (lint-tests CI gate). Add (see #1708).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* feat(#1679): confine install writes within configHome
ADR-1239 Phase B write-confinement: a pure assertDestWithinConfigHome(configDir, destSubpath) rejects a destSubpath that escapes configHome (path traversal / NUL byte) at plan-build time on BOTH the install and uninstall plan paths; surface.applySurface and installOpencodeFamilySkills route through it, and _copyStaged carries a defense-in-depth containment check. Security-load-bearing for the Phase C third-party-descriptor loader.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(#1704): add changeset for destSubpath write-confinement
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(#1704): fix windows path-portability in confinement test
The N1 'accepts a true child subpath' assertion compared against path.join (no drive resolution) while the helper uses path.resolve — on Windows that mismatches the C: drive prefix. Compute the expected via path.resolve to mirror the helper. Windows-CI-only failure (local gsd-test is Mac+Linux).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* feat(#1684): add negotiated host-integration interface module
ADR-1239 Phase A: a pure, additive, no-I/O module exposing PROTOCOL_VERSION, the 8-axis HOST_INTEGRATION_AXES closed vocabulary, the UNDOCUMENTED fail-closed sentinel, negotiateHostCapabilities (effective subset of host-declared and engine-known), a typed degradation ladder, and host-capability profiles.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(#1684): validate and document host-integration axes (16 runtimes)
Extend validateRuntimeBody to validate the 8 hostIntegration axes (closed enums + undocumented sentinel + dispatch struct + reserved-key guards) and the widened runtime vocabulary; author a documentation-sourced hostIntegration block in all 16 runtime descriptors; regenerate the registry. Every per-CLI value is documented (cited) or the explicit undocumented sentinel.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(#1684): add host-integration capability matrix and adr amendment
New per-CLI, per-axis citation reference (value/source/evidence for all 16 CLIs); ADR-1239 Phase-A-implemented amendment; CONTEXT.md glossary seam entry.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(#1684): harden dispatch negotiation edge cases
Code-review hardening: treat NaN/Infinity maxDepth as missing (fail-closed, +warning); reset nested/background when namedDispatch collapses to false (struct consistency); SAFE_DEFAULTS dispatch floor to read-only; warn on non-finite protocolVersion; symmetric undocumented warnings for dispatch fields. Pure module — no consumers; behaviour fail-closed throughout.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(#1684): register host-integration.cjs in lint-ignore and inventory
New tsc-generated bin/lib artifact: add to the eslint ignore list (ADR-457 — lint the .cts source), regenerate docs/INVENTORY-MANIFEST.json, and add the docs/INVENTORY.md CLI-modules row. Fixes the 3 gsd-test failures (551-eslint-bin-lib-coverage x2 + inventory-manifest-sync).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(#1684): add changeset fragment for host-integration interface
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(#1684): add how-to for sourcing a host's integration axes
Diataxis how-to guide for adding/updating a host's runtime.hostIntegration axes from authoritative docs, the undocumented-sentinel rule, validation, and extending the closed vocabulary. Completes the Step-5 doc quadrants (reference + explanation + how-to). Indexed in docs/README.md.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Address maintainer re-review nits:
- Changeset body no longer claims install-exports resolution (Failure 2 is
now handled upstream by #1511 / ADR-1508 Phase 2). This PR delivers the
.gsd-source marker writer, which fixes list/status (Failure 1).
- Add negative tests locking the `runtime === 'claude' && isGlobal` write
guard: a non-claude global install and a claude *local* install both
provision no marker. Verified must-fail-first against a relaxed guard.
* docs(#1650): fix stale opencode install-path claim in core settings
* feat(#1688): warn on stale model bake for static-frontmatter runtimes
* chore(#1688): backfill changeset pr field with real PR number
* test(#1688): make resolveAgentDir assertions use path.join for windows
* docs(#1688): codify windows path-literal-in-assert anti-pattern + align test
Append a dated amendment to ADR-1508 (append-only convention) recording that
the Runtime Artifact Conversion Module decision is implemented on next: the
content-rewrite engine + walkers + computePathPrefix live behind the deep seam
rewriteStagedSkillBodies/rewriteStagedCommandBodies, the getInstallExports relay
and GSD_TEST_MODE require are deleted, and CONTEXT.md marks the module SHIPPED.
Status stays Accepted (no Implemented status per docs/adr/README.md). Two
deferred follow-ups are tracked as sub-issues of the epic (neither a blocker):
- #1675 dedup convertClaudeToAugmentMarkdown family
- #1676 fast-check property test ($HOME-collapse + idempotency)
Delivery verified by a Codex (gpt-5.4, high) read-only review against the
epic's stated deliverables, cross-checked against the indexed code graph.
Closes#1507
Delivers the property coverage promised in #1511's test scope but not landed
(follow-up #1676, epic #1507 / ADR-1508). Adds
tests/enh-1676-path-prefix-collapse-idempotency.property.test.cjs covering:
(A) $HOME-collapse invariant for _computePathPrefix — global-under-home
projects to $HOME/<suffix>/ (exact equality, not substring, so short
homes like /root or /a do not false-positive); opencode is the
documented exception (absolute form, never $HOME).
(B) backslash->posix invariance (#1615 Windows path-leak fix).
(C) path-rewrite idempotency for _applyRuntimeRewrites across the
path-rewriting runtimes (f(f(c)) === f(c); attribution held at
undefined to isolate the path axis). Non-vacuous: a sanity assertion
proves the first pass actually rewrites the seed ~/.claude/ refs.
Pure test addition — no production code changed. Uses the shared
fast-check-setup (seed=42, numRuns=200). Closes#1676
bin/install.js held byte-identical duplicate definitions of the augment
converter family (convertSlashCommandsToAugmentSkillMentions,
convertClaudeToAugmentMarkdown, getAugmentSkillAdapterHeader,
convertClaudeCommandToAugmentSkill, convertClaudeAgentToAugmentAgent) that
already exist canonically in src/runtime-artifact-conversion.cts (generated
to gsd-core/bin/lib/runtime-artifact-conversion.cjs). Deferred Phase 1->2
cleanup tracked in #1675 (epic #1507 / ADR-1508).
Deleted the five local copies; install.js now binds the three PUBLIC
converters from runtimeArtifactConversion (same pattern as getDirName /
processAttribution in #1510). The two private helpers live only in the
conversion module now. module.exports preserved (re-exported).
Behavior-preserving: four converters byte-identical; the fifth
(convertClaudeAgentToAugmentAgent) differed only by an inert let->const
(variable never reassigned). Extends the DEFECT.GENERATIVE-FIX
reference-identity parity guard in enh-1511 to assert single-sourcing.
Closes#1675
Phase 0 of the Dynamic Context Management epic (#1671): the ADR (documentation) plus a non-shipping reference example under examples/dynamic-context-management/ — the Option-E predicate fact-store parser/selector, a self-contained --check/--write/--select index generator, a sample index, and a runnable demo.
The example is intentionally excluded from the build (src/->bin/lib/), the npm package files[], the installer, and the CI test suite (tests/) — nothing here is compiled into or installed with GSD. Production implementation lands in a later phase.
Resolves#1672
Refs #1671
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
The auto-backmerge workflow runs sync-next-version.cjs which calls npm version,
triggering the version lifecycle hook (gen-capability-registry.cjs). That script
requires the compiled capability-validator.cjs (a build:lib artifact), which was
never built on the backmerge runner — causing the backmerge to fail with
'capability validation failed — registry not written'. Add the build step.
* fix(#1520): randomize mktemp temp paths on BSD/macOS (XXXXXX must be path-final)
BSD/macOS mktemp only substitutes the XXXXXX template when it is the final
path component. Templates like `...-XXXXXX.json` / `gsd-pr-body.XXXXXX.md`
return a LITERAL `XXXXXX` path (no randomization) on macOS, so concurrent
workflow runs collide on the same temp manifest/body file — one run can
overwrite or consume another's. Reproduced on macOS: the second call to the
suffixed template fails `mkstemp: File exists`.
Fix: use a suffixless `XXXXXX` template (so it IS the final component), then
rename to add the intended extension — portable across BSD + GNU userlands,
no GNU-only `--suffix` flag. Empty-file-then-write semantics are preserved at
every site.
Affected workflow temp files:
- execute-phase.md: gsd-worktree-wave-*.json (wave worktree manifest)
- quick.md: gsd-quick-worktree-*.json
- spec-phase.md: edge-probe-reqs-*.json
- ship.md: gsd-pr-body-*.md
- profile-user.md: gsd-profile-answers-*.json, gsd-profile-analysis-*.json
The execute-phase.md edit uses a compact intermediate var + trailing comment
to stay under the ADR-857 phase-6 size ceiling (93166); regenerated the
workflow size baseline accordingly. Validated on macOS: 20 concurrent calls
yield 20 unique randomized paths.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* chore(#1520): add changeset fragment (Fixed)
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* test(#1520): add fail-first workflow-prose guard for mktemp XXXXXX suffix
Repo-wide scan of gsd-core/workflows/**/*.md that fails on any mktemp
template whose XXXXXX run is followed by a filename suffix (the BSD/macOS
non-randomizing form). Fails on the six pre-fix instances and passes on
the fix, and locks the copy-paste-prone idiom out of future workflows.
Mirrors the bug-637 hardcoded-$HOME workflow guard.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* test(#1520): rename regression test to fix- prefix (regression-test-names lint)
New tests/bug-NNNN-*.test.cjs files are banned by the lint-regression-test-names
ratchet; use the fix- prefix (matches the fix-1445 precedent).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* test(#1520): add issue ref to allow-test-rule exemption (ADR-456 lint)
lint-allow-test-rule-refs requires every new `allow-test-rule:` comment to
carry a #NNN reference (don't allowlist). Add (#1520) to the source-text
exemption.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(#1520): abort touched mktemp chains on failure (|| exit 1)
Per review: the VAR=$(mktemp …) && mv … && VAR=… chains dropped the issue's
suggested failure guard. If mktemp fails, $VAR is empty and the subsequent
mv/write lands on an unintended relative path. Add `|| exit 1` to all six
touched chains so a mktemp failure aborts the snippet. Regenerated the
workflow size baseline for the slightly longer lines.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* chore(#1520): rebase onto next — regen size baseline + describe rename
Resolve the workflow-size-baseline.json conflict from next advancing by
regenerating from the current workflow sizes. Also rename the test describe
from `bug #1520` to `#1520` (the file uses the fix- prefix) per review nit.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(#1520): harmonize profile-user mktemp to ${TMPDIR:-/tmp} (review nit)
The two profile-user.md temp sites this PR already rewrites kept a hardcoded
/tmp while the four sibling workflows use ${TMPDIR:-/tmp}. Harmonize for
consistency and macOS-correctness (some sandboxes have no writable /tmp).
Regenerated the workflow size baseline.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* chore(#1520): regen size baseline after rebase onto next
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Tom Boucher <trekkie@nomorestars.com>
* feat(#323): fish-shell support in post-install PATH suggestion
Two additive changes to the post-install PATH-suggestion seam, both scoped
to existing functions.
A. Projection: add a fish entry to the persist-mode shell-action list in
projectPathActionProjection() (src/shell-command-projection.cts). fish has
no `export`/`$PATH`-list syntax, so the existing zsh/bash `export PATH=...`
commands are inert when pasted. The new entry emits the fish-native
`fish_add_path '<dir>'` (fish 3.2+, persists via the universal-variable
store, de-duplicating). The directory is single-quoted with the same POSIX
literal escaping as the zsh/bash siblings; verified round-tripping through
real fish 3.7.0 for paths containing quotes, spaces, `$`, `*`, backticks
and unicode.
B. Detection: add homePathCoveredByFishConfig() in bin/install.js, called
from maybeSuggestPathExport() alongside homePathCoveredByRc(). fish does
not use sh-style `export PATH=` rc files, so a fish user whose
fish_user_paths already covers the global bin would otherwise get a
false-positive "not on your PATH" warning on every install. Two
side-effect-free detection routes (no fish subprocess):
1. The universal-variable store (~/.config/fish/fish_variables). fish
serializes this with `full_escape`: every byte outside [A-Za-z0-9/_]
becomes `\xHH` (space -> \x20, `-` -> \x2d, `.` -> \x2e, `$` -> \x24,
unicode -> \uXXXX) and list elements are joined by the literal 4-char
token `\x1e` (NOT a raw 0x1e byte). The detector splits on `\x1e`,
decodes the escapes, then compares each as an absolute literal — a
decoded `$` is part of the directory name, not an unexpanded variable.
Verified against real fish 3.7.0 output.
2. config.fish (`fish_add_path`, `set -gx PATH`, `set -Ux fish_user_paths`)
— plain shell tokens: HOME forms ($HOME/${HOME}/~) are expanded and a
token still holding `$` (e.g. `$PATH`, `$fish_user_paths`) is skipped.
Honours $XDG_CONFIG_HOME and always also checks ~/.config/fish.
No behaviour change for bash/zsh/PowerShell/cmd/Git-Bash users: their entries
and command strings are unchanged; the fish entry is additive and the fish
detector only narrows the set of cases that warn.
Tests: update the projection length assertion (2 -> 3) and fish escaping in
bug-3441; add fish detection + suppression cases in install-path-detection
(uvar store with real fish escaping, dot/hyphen/space/$-literal decode
regressions, config.fish routes, commented-out, relative-segment guard,
unreadable-file fault injection, suppression and emission via
maybeSuggestPathExport).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* chore(changeset): add Changed fragment for #323 fish PATH support (#727)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(#323): address review — action-only fish docs, decoder property test, win32 guard
Addresses @trek-e's review on #727:
- docs (blocker): keep the how-to action-only (Diátaxis). Drop the
`# fish — persists via …` comment and the internal-mechanism clause
naming fish_variables/config.fish; leave one command + the exec-fish
directive.
- tests (minor): extract decodeFishUniversalValue to a pure, exported
module function and add fast-check round-trip properties
(decode(fishEscape(p)) === p over arbitrary unicode, abs-path variant,
totality). Consolidated into install-path-detection.test.cjs to respect
the install test-file-count ratchet.
- tests (follow-up): port #721's win32 negative-projection test (no fish
action on win32; persist projection is PowerShell/cmd.exe/Git Bash).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(#323): address review — drop unused 'after' import, clarify escaping comment
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Tom Boucher <trekkie@nomorestars.com>
* fix(#1577): isolate WebFetch/WebSearch ingress + opt-in injection blocking
Split A of #1573 (security-critical). Scans WebFetch/WebSearch output (the
largest untrusted channel) in gsd-read-injection-scanner; shared
untrusted-input-boundary reference @-included by the 8 ingest agents
(randomized per-wrap delimiters, in-prompt self-scan guard, task-anchoring);
opt-in security.injection_blocking (default advisory — non-breaking).
arXiv: 2506.05739 (PPA), 2507.15219 (PromptArmor), 2504.20472 (Referencing), 2503.00061 (defense-in-depth).
* fix(#1577): address review — honest blocking docs, config key, ADR, property test, revert localized
- A1: rewrote the opt-in-blocking doc + Security changeset honestly — the PostToolUse hook is a
circuit-breaker (halts the agent's next step), NOT a redactor; it does not scrub content already
in the transcript. The prompt-level data/instruction boundary is the primary control.
- A2: registered security.injection_blocking in the config schema + defaults manifests (default
false) + an e2e config-roundtrip test; the dotted setter writes the nested shape the hook reads.
- A3: reverted the 4 hand-edited localized security-model.md (canonical EN only, per convention).
- A5: ADR-1577 (untrusted-input boundary + opt-in blocking; redaction-vs-circuit-breaker rationale).
- A6: property test — scanner never crashes / only emits valid JSON on unicode/large/malformed input.
- Also: inventory (untrusted-input-boundary.md) + agent-size baseline (8 ingest agents) +
drift-guard matcher update (Read -> Read|WebFetch|WebSearch). A7 (content<20 early-exit) left as
the noted pre-existing follow-up.
* fix(#1577): allowlist untrusted-input-boundary.md in injection-scan CI gate
The new reference quotes injection phrases ('ignore previous instructions',
'you are now…') as examples agents must NOT comply with, tripping the repo's
own prompt-injection-scan.sh diff gate (the standalone 'security' CI job, red
on HEAD). Allowlist it alongside the other security docs (security-model.md,
TEST-EXAMPLES.md) that legitimately demonstrate injection patterns. The JS
scanner test doesn't scan references/, so only the shell gate needed it.
Verified: scan --diff origin/next -> 0 findings; scanner JS test 15/15.
* fix(#1577): cover AC #2's gsd-ui-researcher + gsd-assumptions-analyzer
trek-e Major 1: the @-included set dropped two AC #2 agents. Restore them so
no named web-ingress agent is uncovered, keeping the two justified additions
(gsd-ai-researcher, gsd-domain-researcher). Final set = AC's 8 + 2 = 10.
- gsd-ui-researcher carries the full WebSearch/WebFetch + MCP-fetch toolset.
- gsd-assumptions-analyzer reads 5-15 codebase source files (external/source-
document ingress per the boundary), though it has no web tools.
INGEST_AGENTS in the isolation test now asserts all 10; size baselines
regenerated (+60 bytes each, both well under the DEFAULT cap); changeset
reworded 8 -> 10.
Verified: untrusted-input-isolation 14/14; agent-size-budget 39/39.
* docs(#1577): document security.injection_blocking + boundary seam
trek-e Major 2 + Minor:
- docs/CONFIGURATION.md: add the top-level security.injection_blocking key to
the Full Schema and a Security Settings subsection, distinguishing it from
the workflow.security_* namespace; honest circuit-breaker-not-redactor
framing matching ADR-1577 / security-model.
- CONTEXT.md: add the 'Untrusted-input boundary' seam glossary entry.
Verified: lint:docs ok; config-field-docs + contributor-standards green.
* test(#1577): make read-injection property test git-text, not binary
trek-e nit (and more): the file embedded a raw U+FFFF AND a raw NUL byte as
degenerate-edge inputs. The NUL is what actually made git classify it binary
(git binary = NUL in first 8K). Replace both with text-safe escapes that keep
the identical runtime values: '\\x00' and String.fromCodePoint(0xFFFF). File
now diffs/blames line-by-line.
Verified: property test 2/2; no NUL/raw-noncharacter bytes remain.
* docs(#1577): align untrusted boundary docs
Name all 10 ingress agents in INVENTORY/security-model and allowlist the intentional read-injection property corpus for the prompt-injection scanner.
* docs(#1577): align ADR ingest agent count
Update ADR-1577 from 8 to 10 ingest agents so it matches the actual boundary include set and the rest of the docs.
---------
Co-authored-by: Tom Boucher <trekkie@nomorestars.com>
