Conversation
Part A of the concept-debt reduction pass. Ran the hard verification
protocol across crates/, tools/, scripts/, prototypes/, templates/,
assets/, and root-level scripts. Result: zero high-confidence
deletions. This is a legitimate outcome of the protocol, not a skipped
step — recorded here with the evidence so the next sweep does not
re-walk the same ground.
Candidates investigated and why each was kept:
- cargo build's ~100 dead-code warnings on crates/code-intel-cli: this
repo's own AGENTS.md already documents these as a known
false-positive source ("almost none of them are removable... do not
treat the count as a cleanup backlog"). The crate has no lib.rs;
tests/*.rs integration binaries re-include source files via
`#[path = "..."] mod x;`, so each instantiation only exercises a
slice of a file and warns on the rest. Verified by hand on the 4
warning-heaviest files: admissibility.rs, authority.rs, and
method_catalog.rs are each #[path]-included by multiple live
tests/*.rs binaries. graph.rs's flagged `Options`/`run` have a
direct, unconditional call site from `main()` -> `run()` ->
`cmd_graph()` -> `graph::run(...)`.
- tools/normalize_compete_score.py: referenced by
legacy/Invoke-CompeteProjectScore.ps1 — not zero inbound references.
- crates/code-nexus-lite (no .rs files, only README.md): its own
README documents a deliberate demotion pending an upstream CVE fix
in iii-sdk's opentelemetry_sdk dependency chain, and it is exercised
by .github/workflows/skill-check.yml's link-check job.
- prototypes/session-observability: a scoped, self-documented
throwaway prototype (own IDEA.md) with an explicit "Decision after
trial" gate not yet reached; also a listed repo entry point.
- Run-ScopedRepowiseDocs.py, test-scoped-repowise-validator.py (repo
root): both invoked directly by .github/workflows/ci.yml.
- templates/*.md (dependency-audit, idea-file, minimax-deploy-checklist,
understanding-report — zero literal inbound references each): each
is the generating template for a still-current pattern
(docs/plans/*-idea.md matches templates/idea-file.md's structure
field-for-field) or documents a still-real legacy script/artifact
path; none reference a nonexistent file.
- docs/ponytail-gain-ledger.md (the repo's own tracked minimalism
backlog) independently corroborates the crates/ findings:
code-nexus-lite's workspace demotion (PG-003) and nested Cargo.lock
removal (PG-004) are already-harvested gains, and
legacy/tools/sentrux-shim plus the Sentrux V overlay are explicitly
"Retained After Audit" (PK-001, PK-002). Note for scope honesty: the
same ledger separately flags pipeline.config.json (PG-001) and
legacy/Invoke-SentruxAgentTool.ps1 (PG-006) as open gains outside
Part A's crates/tools/scripts/prototypes scope — those are the
ledger's own open items, not cleared by this sweep.
- No git-log evidence of a superseded/duplicate module left behind
(checked sentrux*.rs, doctor*.rs, artifact*.rs clusters against
retire/drop/supersede commit history).
Before: investigated every candidate in Part A's stated scope.
After: 0 files removed, 0 lines removed.
Verified `cargo build -p code-intel --release --locked` and
`cargo test -p code-intel --locked` both exit 0 on this tree (212
lib-slice tests plus all integration-test binaries passing) prior to
this investigation, with no crates/ source changes made.
Refs #55 #104
…s seeded Part B (docs lifecycle sweep) + Part C (.out-of-scope registry). Part B: inventoried all 123 tracked files under docs/ (excluding README/CHANGELOG/AGENTS/CONTEXT at repo root, which are out of scope by spec). Classified and recorded every file in the new docs/INVENTORY.md (machine-first table: path, class, disposition, reason). Every one-shot disposition was confirmed against a primary source before moving, not from filename pattern alone. Before: 123 docs files (14 ADR permanent-and-untouched, 2 already archived from a prior sweep, 107 candidates for this classification). After: 17 moved to docs/archive/ via git mv (history preserved); 104 kept in place as living-contract (14 ADR + 90 actively-referenced docs); the 2 pre-existing archive entries untouched; 0 classified stale (no doc found describing wholly nonexistent paths/symbols). Archived (16 of 17 docs/plans/*-idea.md — each individually verified IMPLEMENTED/shipped via its status header plus a corresponding real doc, source module, test, or orchestration/internalization/*.json record; kept docs/plans/adr-0010-execution-plan.md in place since its own header says "proposed, not implemented" and it is cited by 3 other live files): automatic-pr-one-command-orchestration-idea, compete-project-score-idea, four-blind-spots-closure-idea, language-adapter-acceptance-standard-idea, model-independent-pipeline-completion-idea, multi-agent-merge-queue-idea, multi-agent-workspace-governance-idea, pon-multilanguage-code-evidence-idea, pon-parity-floor-idea, pon-project-conformance-mapping-idea, python314-pon-development-lane-idea, sentrux-rust-analysis-kernel-idea, session-evidence-adapter-idea, three-stage-project-acceptance-idea, understanding-quadrant-complexity-idea, v0.3.0-release-bootstrap-and-layout-idea. Archived (1 of 8 docs/compatibility-retire-*.md): only compatibility-retire-doctor-wrapper-branch.md has orchestration/retirements/e09-doctor-wrapper/status.json with retired=true, deletionExecuted=true. This task's own prompt text assumed the whole compatibility-retire-* category was "already executed" — primary-source verification (status.json for all 8 packets) found the other 7 are still decision=blocked, deletionExecuted=false: active, open retirement tickets, not completed campaigns. Kept in place, not archived. All 3 docs/ps1-exit/*.md were kept for the parallel reason: each is directly cited from live Rust source (hospital_score.rs, execution_policy/inputs.rs) or is the companion doc to the CI-invoked PS1-vs-Rust parity harness (test-ps1-rust-parity.ps1, run by .github/workflows/parity-observe.yml). Known accepted side effect: docs/final-commitment-reconciliation.md still cites compatibility-retire-doctor-wrapper-branch.md at its old path in one table cell. Left as-is — updating cross-references in surviving living docs is out of this sweep's stated scope, and git history plus docs/INVENTORY.md make the new location discoverable. Part C: created .out-of-scope/ with the three seeded non-goal files from the spec (unauthenticated-local-http-daemon.md, human-first-artifacts-as-source.md, slash-invoked-product-shape.md), each stating the non-goal, why, and what to do instead. Gates run after these changes (docs-only; no crates/ touched, so the build/test green recorded in the prior commit still holds): the authoritative self-scan (`run execute --manifest orchestration/integrations.json`) exit 0, every node succeeded/pass, zero domain/process failures. `eval/harness.py` exit 0, reproducing the committed baseline's Arm A 6/12 / Arm B 9/12 coverage numbers unchanged (eval/ working tree restored to its pre-run state afterward — the harness's own baseline-rotation behavior is not part of this sweep's deliverable). Refs #55 #104
q06 (keywords "pipeline_root" + "capability_inventory") silently flipped
Arm B from covered to golden_file_outside_top_cap between the committed
baseline and a fresh re-run: eval/questions.json contains those two
keywords 11 times in its own JSON text (verified), so it -- and
eval/BASELINE.md's rendered per-question summary -- ranked in Arm B's
naive keyword search and diluted the 5-file cap, pushing the real golden
file (capability_inventory.rs) out. Root cause is the benchmark reading
its own corpus and outputs, not a real regression in either arm's logic.
Fix: eval/arms.py gains EXCLUDED_CORPUS_PREFIXES ("eval/", "docs/archive/",
".out-of-scope/") and an `is_excluded_from_corpus()` predicate, applied at
enumeration (git_ls_files) and at each arm's point of use (arm_b_answer's
ranking loop, search_kind's hit collection) so the exclusion holds
regardless of how file_universe/artifact data was built. docs/archive/
(one-shot docs already classified non-current) and .out-of-scope/ (a
non-goal registry) are excluded on the same "never a valid answer source"
principle, per this follow-up's own scope.
Stdlib-only, deterministic; no Rust changed. `python eval/test_harness.py`
passes (26/26, was 23/23): added CorpusExclusionTests with a fixture file
under an excluded dir (eval/fixtures/repo/eval/questions.json) that
deliberately outscores the real golden files, proving it never reaches
either arm's files_touched despite the higher score.
Refs #93 #104
Rotate eval/baseline-8a79814.json (captured before the corpus-exclusion fix) to eval/baseline-d889b39.json (HEAD at the fix commit), with BASELINE.md regenerated to match. Fresh run after the fix: Arm A 6/12 covered (14,959,835 bytes), Arm B 10/12 covered (3,414,038 bytes) -- q06 is covered:true on both arms again, matching the pre-pollution baseline's coverage state exactly (byte counts differ only by the amount the repo's own content has grown since 8a79814, e.g. docs/INVENTORY.md, .out-of-scope/). Ran twice; diffed both outputs with the `meta` object excluded (timestamps, the nonce-bearing pipeline run directory, and wall-clock timing live only there per harness.py's own determinism contract): byte-identical. Refs #93
|
Caution The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased. |
|
🔒 Repowise is not analyzing this repository The PR bot is free on public repositories. This one is private, which needs a Pro plan. |
ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (27)
📝 WalkthroughSummary by CodeRabbit
WalkthroughThe change adds documentation and archive records. It also excludes ChangesDocumentation records
Evaluation corpus exclusions
Estimated code review effort: 3 (Moderate) | ~20 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing Touches 💡 1📝 Generate docstrings 💡
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Code Intel change risk
Top signals
revspec: |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 8
🧹 Nitpick comments (2)
docs/archive/plans/compete-project-score-idea.md (2)
13-15: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winUse a repository-neutral target path.
Replace
D:/projects/_tools/code-intel-pipelinewith<repo>or a repository-relative identifier. The current value makes this archive record depend on one developer's filesystem layout.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/archive/plans/compete-project-score-idea.md` around lines 13 - 15, Replace the developer-specific absolute path in the archive plan’s “Path” entry with the repository-neutral placeholder <repo> or an equivalent repository-relative identifier, while leaving the branch and current-state entries unchanged.
4-4: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winPin the external source revision.
The
Sourcefield points to a moving repository URL. Record an immutable commit or tag for thecompeterevision used bybuild_report.py. This keeps the archived scoring contract reproducible.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/archive/plans/compete-project-score-idea.md` at line 4, Update the Source field in compete-project-score-idea.md to reference the specific immutable commit hash or release tag of the compete repository used by build_report.py, replacing the moving repository URL while preserving the archived scoring contract.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.out-of-scope/slash-invoked-product-shape.md:
- Around line 20-21: Update the prose in slash-invoked-product-shape.md so the
issue reference is on the same line as the surrounding text, using “Refs issue
`#105`.” rather than starting a wrapped line with “#105.”; preserve the existing
meaning and wording otherwise.
In `@docs/archive/plans/automatic-pr-one-command-orchestration-idea.md`:
- Around line 7-10: Replace the PowerShell-first design with the compiled
code-intel CLI as the primary interface: update
docs/archive/plans/automatic-pr-one-command-orchestration-idea.md lines 7-10 to
specify a CLI entrypoint; update
docs/archive/plans/language-adapter-acceptance-standard-idea.md lines 22 and
38-41 so acceptance uses the compiled CLI and PowerShell is compatibility-only;
update docs/archive/plans/multi-agent-merge-queue-idea.md lines 48-52 to use the
CLI landing-adapter surface; and update
docs/archive/plans/multi-agent-workspace-governance-idea.md lines 26-31 and
37-41 to remove PowerShell-first constraints and make the CLI the preflight
facade. Lead new documentation and command examples with code-intel, mentioning
PowerShell only for existing compatibility paths.
In `@docs/archive/plans/four-blind-spots-closure-idea.md`:
- Line 30: Update the “Runtime/CI ingestion” sentence to use the compound
modifier “file/request-based” instead of “file/request based,” preserving the
rest of the sentence unchanged.
In `@docs/archive/plans/multi-agent-workspace-governance-idea.md`:
- Line 2: Resolve the contradiction between the IMPLEMENTED status and the
unchecked report-pack criterion in the plan: either complete and check the
requirement that the pipeline emits summary.md, report.json, and
understanding.md, then retain IMPLEMENTED, or mark that criterion deferred/not
applicable and document the scope exclusion.
In `@docs/archive/plans/three-stage-project-acceptance-idea.md`:
- Around line 18-19: Update the acceptance plan to lead with the compiled
code-intel CLI for setup and execution, and identify
legacy/Invoke-CodeIntelAcceptance.ps1 and legacy/install-code-intel-pipeline.ps1
as compatibility-only paths. Revise the related sections around the checked
items and lines 35-38 so command examples and new guidance use the CLI first,
mentioning PowerShell only for existing compatibility support.
In `@docs/archive/plans/v0.3.0-release-bootstrap-and-layout-idea.md`:
- Around line 1-4: Update the archived plan header in “v0.3.0 Release Bootstrap
and PowerShell Layout” from ACTIVE to the inventory’s archived/completed status,
and mark the release-criteria checklist as completed to reflect the shipped
v0.3.0 outcome. Keep the file in docs/archive/ only if its metadata and
checklist consistently show completion.
In `@eval/arms.py`:
- Around line 67-72: Update is_excluded_from_corpus to canonicalize
repository-relative paths by normalizing dot segments, reject absolute paths and
paths escaping the repository, then compare the canonical path against
EXCLUDED_CORPUS_PREFIXES. Extend CorpusExclusionTests with ./, ../, and
absolute-path cases while preserving exclusion checks for valid normalized
paths.
In `@eval/BASELINE.md`:
- Line 3: Update the baseline renderer in eval/harness.py so generated command
guidance leads with the compiled code-intel CLI instead of python
eval/harness.py --render, then regenerate eval/BASELINE.md with the renderer and
preserve its generated-file marker.
---
Nitpick comments:
In `@docs/archive/plans/compete-project-score-idea.md`:
- Around line 13-15: Replace the developer-specific absolute path in the archive
plan’s “Path” entry with the repository-neutral placeholder <repo> or an
equivalent repository-relative identifier, while leaving the branch and
current-state entries unchanged.
- Line 4: Update the Source field in compete-project-score-idea.md to reference
the specific immutable commit hash or release tag of the compete repository used
by build_report.py, replacing the moving repository URL while preserving the
archived scoring contract.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro Plus
Run ID: 0b7a35fb-eadd-413e-ad24-cd2b165e7cb5
📒 Files selected for processing (27)
.out-of-scope/human-first-artifacts-as-source.md.out-of-scope/slash-invoked-product-shape.md.out-of-scope/unauthenticated-local-http-daemon.mddocs/INVENTORY.mddocs/archive/compatibility-retire-doctor-wrapper-branch.mddocs/archive/plans/automatic-pr-one-command-orchestration-idea.mddocs/archive/plans/compete-project-score-idea.mddocs/archive/plans/four-blind-spots-closure-idea.mddocs/archive/plans/language-adapter-acceptance-standard-idea.mddocs/archive/plans/model-independent-pipeline-completion-idea.mddocs/archive/plans/multi-agent-merge-queue-idea.mddocs/archive/plans/multi-agent-workspace-governance-idea.mddocs/archive/plans/pon-multilanguage-code-evidence-idea.mddocs/archive/plans/pon-parity-floor-idea.mddocs/archive/plans/pon-project-conformance-mapping-idea.mddocs/archive/plans/python314-pon-development-lane-idea.mddocs/archive/plans/sentrux-rust-analysis-kernel-idea.mddocs/archive/plans/session-evidence-adapter-idea.mddocs/archive/plans/three-stage-project-acceptance-idea.mddocs/archive/plans/understanding-quadrant-complexity-idea.mddocs/archive/plans/v0.3.0-release-bootstrap-and-layout-idea.mdeval/BASELINE.mdeval/arms.pyeval/baseline-d889b39.jsoneval/fixtures/repo/eval/questions.jsoneval/harness.pyeval/test_harness.py
Comment on lines
+20
to
+21
| that evidence would make the pipeline's output as easy to skip as any other optional tool. Refs | ||
| #105. |
There was a problem hiding this comment.
📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win
Keep the issue reference on the prose line.
Line 21 begins with #105. after a line wrap. Markdownlint reports a missing space after the heading marker. Write Refs issue #105. on one line to avoid malformed Markdown.
🧰 Tools
🪛 markdownlint-cli2 (0.23.1)
[warning] 21-21: No space after hash on atx style heading
(MD018, no-missing-space-atx)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In @.out-of-scope/slash-invoked-product-shape.md around lines 20 - 21, Update
the prose in slash-invoked-product-shape.md so the issue reference is on the
same line as the surrounding text, using “Refs issue `#105`.” rather than starting
a wrapped line with “#105.”; preserve the existing meaning and wording
otherwise.
Source: Linters/SAST tools
Comment on lines
+67
to
+72
| def is_excluded_from_corpus(rel_path: str) -> bool: | ||
| """True if `rel_path` (repo-relative) is the benchmark's own corpus, | ||
| its own output, or archived/non-goal material -- never a valid answer | ||
| source for either arm.""" | ||
| rel = posix(rel_path) | ||
| return any(rel.startswith(prefix) for prefix in EXCLUDED_CORPUS_PREFIXES) |
There was a problem hiding this comment.
🎯 Functional Correctness | 🟠 Major | ⚡ Quick win
Normalize dot segments before checking exclusion prefixes.
posix() only converts separators. It does not normalize ./eval/questions.json or docs/tmp/../archive/plan.md. These paths bypass the prefix check, and repo_root / rel_path can then read excluded corpus content.
Canonicalize repository-relative paths before comparison. Reject absolute paths and paths that escape the repository. Add these cases to CorpusExclusionTests.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@eval/arms.py` around lines 67 - 72, Update is_excluded_from_corpus to
canonicalize repository-relative paths by normalizing dot segments, reject
absolute paths and paths escaping the repository, then compare the canonical
path against EXCLUDED_CORPUS_PREFIXES. Extend CorpusExclusionTests with ./, ../,
and absolute-path cases while preserving exclusion checks for valid normalized
paths.
| # Eval v1 baseline -- artifact-guided (A) vs naive (B) | ||
|
|
||
| Generated by `python eval/harness.py --render` from `eval/baseline-8a79814.json`. Do not hand-edit; re-run harness.py instead. | ||
| Generated by `python eval/harness.py --render` from `eval/baseline-d889b39.json`. Do not hand-edit; re-run harness.py instead. |
There was a problem hiding this comment.
📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win
Lead with the compiled code-intel CLI.
Line 3 presents python eval/harness.py --render as the generation command. Update the baseline renderer so its generated command guidance starts with the compiled code-intel CLI, then regenerate this file.
As per coding guidelines, new documentation and command examples should lead with the compiled code-intel CLI.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@eval/BASELINE.md` at line 3, Update the baseline renderer in eval/harness.py
so generated command guidance leads with the compiled code-intel CLI instead of
python eval/harness.py --render, then regenerate eval/BASELINE.md with the
renderer and preserve its generated-file marker.
Source: Coding guidelines
There was a problem hiding this comment.
Caution
Inline review comments failed to post. This is likely due to GitHub's internal server error or limits when posting large numbers of comments. If you are seeing this consistently it is likely a permissions issue. Please check "Moderation" -> "Code review limits" under your organization settings.
Actionable comments posted: 8
🧹 Nitpick comments (2)
docs/archive/plans/compete-project-score-idea.md (2)
13-15: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winUse a repository-neutral target path.
Replace
D:/projects/_tools/code-intel-pipelinewith<repo>or a repository-relative identifier. The current value makes this archive record depend on one developer's filesystem layout.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/archive/plans/compete-project-score-idea.md` around lines 13 - 15, Replace the developer-specific absolute path in the archive plan’s “Path” entry with the repository-neutral placeholder <repo> or an equivalent repository-relative identifier, while leaving the branch and current-state entries unchanged.
4-4: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winPin the external source revision.
The
Sourcefield points to a moving repository URL. Record an immutable commit or tag for thecompeterevision used bybuild_report.py. This keeps the archived scoring contract reproducible.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/archive/plans/compete-project-score-idea.md` at line 4, Update the Source field in compete-project-score-idea.md to reference the specific immutable commit hash or release tag of the compete repository used by build_report.py, replacing the moving repository URL while preserving the archived scoring contract.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.out-of-scope/slash-invoked-product-shape.md:
- Around line 20-21: Update the prose in slash-invoked-product-shape.md so the
issue reference is on the same line as the surrounding text, using “Refs issue
`#105`.” rather than starting a wrapped line with “#105.”; preserve the existing
meaning and wording otherwise.
In `@docs/archive/plans/automatic-pr-one-command-orchestration-idea.md`:
- Around line 7-10: Replace the PowerShell-first design with the compiled
code-intel CLI as the primary interface: update
docs/archive/plans/automatic-pr-one-command-orchestration-idea.md lines 7-10 to
specify a CLI entrypoint; update
docs/archive/plans/language-adapter-acceptance-standard-idea.md lines 22 and
38-41 so acceptance uses the compiled CLI and PowerShell is compatibility-only;
update docs/archive/plans/multi-agent-merge-queue-idea.md lines 48-52 to use the
CLI landing-adapter surface; and update
docs/archive/plans/multi-agent-workspace-governance-idea.md lines 26-31 and
37-41 to remove PowerShell-first constraints and make the CLI the preflight
facade. Lead new documentation and command examples with code-intel, mentioning
PowerShell only for existing compatibility paths.
In `@docs/archive/plans/four-blind-spots-closure-idea.md`:
- Line 30: Update the “Runtime/CI ingestion” sentence to use the compound
modifier “file/request-based” instead of “file/request based,” preserving the
rest of the sentence unchanged.
In `@docs/archive/plans/multi-agent-workspace-governance-idea.md`:
- Line 2: Resolve the contradiction between the IMPLEMENTED status and the
unchecked report-pack criterion in the plan: either complete and check the
requirement that the pipeline emits summary.md, report.json, and
understanding.md, then retain IMPLEMENTED, or mark that criterion deferred/not
applicable and document the scope exclusion.
In `@docs/archive/plans/three-stage-project-acceptance-idea.md`:
- Around line 18-19: Update the acceptance plan to lead with the compiled
code-intel CLI for setup and execution, and identify
legacy/Invoke-CodeIntelAcceptance.ps1 and legacy/install-code-intel-pipeline.ps1
as compatibility-only paths. Revise the related sections around the checked
items and lines 35-38 so command examples and new guidance use the CLI first,
mentioning PowerShell only for existing compatibility support.
In `@docs/archive/plans/v0.3.0-release-bootstrap-and-layout-idea.md`:
- Around line 1-4: Update the archived plan header in “v0.3.0 Release Bootstrap
and PowerShell Layout” from ACTIVE to the inventory’s archived/completed status,
and mark the release-criteria checklist as completed to reflect the shipped
v0.3.0 outcome. Keep the file in docs/archive/ only if its metadata and
checklist consistently show completion.
In `@eval/arms.py`:
- Around line 67-72: Update is_excluded_from_corpus to canonicalize
repository-relative paths by normalizing dot segments, reject absolute paths and
paths escaping the repository, then compare the canonical path against
EXCLUDED_CORPUS_PREFIXES. Extend CorpusExclusionTests with ./, ../, and
absolute-path cases while preserving exclusion checks for valid normalized
paths.
In `@eval/BASELINE.md`:
- Line 3: Update the baseline renderer in eval/harness.py so generated command
guidance leads with the compiled code-intel CLI instead of python
eval/harness.py --render, then regenerate eval/BASELINE.md with the renderer and
preserve its generated-file marker.
---
Nitpick comments:
In `@docs/archive/plans/compete-project-score-idea.md`:
- Around line 13-15: Replace the developer-specific absolute path in the archive
plan’s “Path” entry with the repository-neutral placeholder <repo> or an
equivalent repository-relative identifier, while leaving the branch and
current-state entries unchanged.
- Line 4: Update the Source field in compete-project-score-idea.md to reference
the specific immutable commit hash or release tag of the compete repository used
by build_report.py, replacing the moving repository URL while preserving the
archived scoring contract.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro Plus
Run ID: 0b7a35fb-eadd-413e-ad24-cd2b165e7cb5
📒 Files selected for processing (27)
.out-of-scope/human-first-artifacts-as-source.md.out-of-scope/slash-invoked-product-shape.md.out-of-scope/unauthenticated-local-http-daemon.mddocs/INVENTORY.mddocs/archive/compatibility-retire-doctor-wrapper-branch.mddocs/archive/plans/automatic-pr-one-command-orchestration-idea.mddocs/archive/plans/compete-project-score-idea.mddocs/archive/plans/four-blind-spots-closure-idea.mddocs/archive/plans/language-adapter-acceptance-standard-idea.mddocs/archive/plans/model-independent-pipeline-completion-idea.mddocs/archive/plans/multi-agent-merge-queue-idea.mddocs/archive/plans/multi-agent-workspace-governance-idea.mddocs/archive/plans/pon-multilanguage-code-evidence-idea.mddocs/archive/plans/pon-parity-floor-idea.mddocs/archive/plans/pon-project-conformance-mapping-idea.mddocs/archive/plans/python314-pon-development-lane-idea.mddocs/archive/plans/sentrux-rust-analysis-kernel-idea.mddocs/archive/plans/session-evidence-adapter-idea.mddocs/archive/plans/three-stage-project-acceptance-idea.mddocs/archive/plans/understanding-quadrant-complexity-idea.mddocs/archive/plans/v0.3.0-release-bootstrap-and-layout-idea.mdeval/BASELINE.mdeval/arms.pyeval/baseline-d889b39.jsoneval/fixtures/repo/eval/questions.jsoneval/harness.pyeval/test_harness.py
🛑 Comments failed to post (5)
docs/archive/plans/automatic-pr-one-command-orchestration-idea.md (1)
7-10: 📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win
Use one CLI-first contract for new adapters and gates.
These documents independently make PowerShell the primary interface for new capabilities. Define the compiled
code-intelCLI first. Keep PowerShell only behind an existing compatibility path.
docs/archive/plans/automatic-pr-one-command-orchestration-idea.md#L7-L10: Replace the new PowerShell entrypoint with the compiled CLI entrypoint.docs/archive/plans/language-adapter-acceptance-standard-idea.md#L22-L22: Make the compiled CLI the acceptance-gate interface.docs/archive/plans/language-adapter-acceptance-standard-idea.md#L38-L41: Make PowerShell validation compatibility-only.docs/archive/plans/multi-agent-merge-queue-idea.md#L48-L52: Make the compiled CLI the landing-adapter surface.docs/archive/plans/multi-agent-workspace-governance-idea.md#L26-L31: Remove the PowerShell-first implementation constraint.docs/archive/plans/multi-agent-workspace-governance-idea.md#L37-L41: Make the compiled CLI the preflight facade.As per coding guidelines, new documentation and command examples should lead with the compiled
code-intelCLI; mention PowerShell only when documenting an existing compatibility path.📍 Affects 4 files
docs/archive/plans/automatic-pr-one-command-orchestration-idea.md#L7-L10(this comment)docs/archive/plans/language-adapter-acceptance-standard-idea.md#L22-L22docs/archive/plans/language-adapter-acceptance-standard-idea.md#L38-L41docs/archive/plans/multi-agent-merge-queue-idea.md#L48-L52docs/archive/plans/multi-agent-workspace-governance-idea.md#L26-L31docs/archive/plans/multi-agent-workspace-governance-idea.md#L37-L41🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/archive/plans/automatic-pr-one-command-orchestration-idea.md` around lines 7 - 10, Replace the PowerShell-first design with the compiled code-intel CLI as the primary interface: update docs/archive/plans/automatic-pr-one-command-orchestration-idea.md lines 7-10 to specify a CLI entrypoint; update docs/archive/plans/language-adapter-acceptance-standard-idea.md lines 22 and 38-41 so acceptance uses the compiled CLI and PowerShell is compatibility-only; update docs/archive/plans/multi-agent-merge-queue-idea.md lines 48-52 to use the CLI landing-adapter surface; and update docs/archive/plans/multi-agent-workspace-governance-idea.md lines 26-31 and 37-41 to remove PowerShell-first constraints and make the CLI the preflight facade. Lead new documentation and command examples with code-intel, mentioning PowerShell only for existing compatibility paths.Source: Coding guidelines
docs/archive/plans/four-blind-spots-closure-idea.md (1)
30-30: 📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win
Use a compound modifier.
Change
file/request basedtofile/request-based.🧰 Tools
🪛 LanguageTool
[grammar] ~30-~30: Use a hyphen to join words.
Context: .... - Runtime/CI ingestion is file/request based and read-only; it does not mutate ...(QB_NEW_EN_HYPHEN)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/archive/plans/four-blind-spots-closure-idea.md` at line 30, Update the “Runtime/CI ingestion” sentence to use the compound modifier “file/request-based” instead of “file/request based,” preserving the rest of the sentence unchanged.Source: Linters/SAST tools
docs/archive/plans/multi-agent-workspace-governance-idea.md (1)
2-2: 📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win
Resolve the implemented-status contradiction.
Status: IMPLEMENTEDconflicts with the unchecked criterion[ ] Pipeline emits summary.md, report.json, and understanding.md.If report-pack generation is outside this slice, mark the criterion as deferred or not applicable and state that scope. Otherwise, update the status only after the criterion is complete.
Also applies to: 17-24
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/archive/plans/multi-agent-workspace-governance-idea.md` at line 2, Resolve the contradiction between the IMPLEMENTED status and the unchecked report-pack criterion in the plan: either complete and check the requirement that the pipeline emits summary.md, report.json, and understanding.md, then retain IMPLEMENTED, or mark that criterion deferred/not applicable and document the scope exclusion.docs/archive/plans/three-stage-project-acceptance-idea.md (1)
18-19: 📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win
Lead with the compiled CLI.
This document presents
legacy/Invoke-CodeIntelAcceptance.ps1as the acceptance entry point andlegacy/install-code-intel-pipeline.ps1as the first setup step. It does not label these paths as compatibility-only. Lead with the compiledcode-intelCLI. Keep PowerShell only as an explicitly documented compatibility path.As per coding guidelines, new documentation and command examples should lead with the compiled
code-intelCLI; mention PowerShell only when documenting an existing compatibility path.Also applies to: 35-38
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/archive/plans/three-stage-project-acceptance-idea.md` around lines 18 - 19, Update the acceptance plan to lead with the compiled code-intel CLI for setup and execution, and identify legacy/Invoke-CodeIntelAcceptance.ps1 and legacy/install-code-intel-pipeline.ps1 as compatibility-only paths. Revise the related sections around the checked items and lines 35-38 so command examples and new guidance use the CLI first, mentioning PowerShell only for existing compatibility support.Source: Coding guidelines
docs/archive/plans/v0.3.0-release-bootstrap-and-layout-idea.md (1)
1-4: 🗄️ Data Integrity & Integration | 🟠 Major | ⚡ Quick win
Align the archive status with the inventory.
docs/INVENTORY.mdrecords this file as archived becausev0.3.0shipped, but the header still saysStatus: ACTIVE. The release criteria also remain unchecked. This contradictory state can make a completed plan appear active. Update the status and checklist to match the verified outcome, or keep the file outsidedocs/archive/.Also applies to: 25-34
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/archive/plans/v0.3.0-release-bootstrap-and-layout-idea.md` around lines 1 - 4, Update the archived plan header in “v0.3.0 Release Bootstrap and PowerShell Layout” from ACTIVE to the inventory’s archived/completed status, and mark the release-criteria checklist as completed to reflect the shipped v0.3.0 outcome. Keep the file in docs/archive/ only if its metadata and checklist consistently show completion.
2233admin
added a commit
that referenced
this pull request
Aug 1, 2026
…116) PR #109 把 docs/compatibility-retire-doctor-wrapper-branch.md 归档（status.json 显示 retired=true），但 Test-FinalCommitmentReconciliation 按字面路径引用它作证据物，windows-build-test-package 因此红。包冻结语义：证据路径永久冻结，与退役状态无关。搬回原路径，INVENTORY 行改判 living-contract 并记规则。 本地验证：test-retirement-packets.ps1 全过（8 packets, 2 audits, final commitment reconciliation PASS）。 Refs #55 #104
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
干什么（收敛处方第一刀，Refs #55 #104）
四个 commit，三件事 + 一个验收时抓到的真 bug：
1. 死代码清扫：0 删除（诚实负结果）
全部候选（crates 告警文件、tools/、prototypes/、根脚本、templates/）逐个 rg 全仓追引用——
#[path]测试内联、活 Rust 源引用、CI 调用、或自文档声明故意保留。AGENTS.md 独立佐证 ~100 条 build 告警是已知#[path]重包含伪影，不是清理欠账。结论：仓库此前的退役战役已经把死肉清干净了；负结果入档，防止后人重复扫。2. 文档生命周期首扫（#104 的人工前哨）
123 篇盘点 → 104 活契约保留，17 篇归档（16 篇
docs/plans/*-idea.md+ 1 篇已完成退役文档），0 stale。关键判断：任务书假设compatibility-retire-*全部已完成——按一手源orchestration/retirements/*/status.json核实，8 个里只有 1 个retired=true，其余 7 个仍 blocked/active，保留不归档。全量机读表在docs/INVENTORY.md。3.
.out-of-scope/非目标注册表（mattpocock/skills 模式）三个种子非目标：无鉴权本地 HTTP 端口 / 人读产物当真源 / 斜杠调用产品形态。往后 scope 回潮先撞这面墙。
4. eval 基准自指污染修复（验收实测抓获）
归档后重跑 eval，Arm B 的 q06 翻 False（10/12→9/12）：朴素臂的关键词搜索命中了
eval/questions.json自己（11 hits），把黄金文件挤出 5 文件帽。基准读自己的语料 = committed 基线与新鲜运行永久打架，正是 #104 禁止的"过时缓存当真源"。修复：两臂搜索空间排除eval/、docs/archive/、.out-of-scope/（新增 3 个排除测试，26/26 绿），基线轮换至可复现态：Arm A 6/12 / Arm B 10/12，双跑逐字节一致。门禁
权威 self-scan exit 0（前后两次）；cargo build/test 全绿；无新巨石文件。
备注：本 PR 的 diff 形状（17 个文档 move 在
--no-renames下算作删+增）会被 pr-gate 高估——属 v1 已知局限，走risk-accepted标签。Refs #55 #104 #93