[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7A0KZ6PWAB5FKA7P502AP1P":3,"skill-same:01M7A0KZ6PWAB5FKA7P502AP1P":44},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":26,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":29,"install":30,"resources":43},"01M7A0KZ6PWAB5FKA7P502AP1P","coderos-source-reuse-auditor",null,"Apply the Source Reuse Acceptance Auditor rubric before opening, and before merging, any PR that copies, adapts, mines, vendors, clean-room-reimplements, sidecar-wraps, or references a public repository -- the acceptance gate that sits above the license\u002Fsecurity boundary, not a substitute for it.","ChristianV997\u002FCoderOS","ChristianV997","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F227661135?v=4","01M7A0CBPF0W3PYCR02GVS7M59","01M7A0CBQ10CQPW99BNMXC5496",1,"MIT","https:\u002F\u002Fgithub.com\u002FChristianV997\u002FCoderOS\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcoderos-source-reuse-auditor\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: coderos-source-reuse-auditor\ndescription: Apply the Source Reuse Acceptance Auditor rubric before opening, and before merging, any PR that copies, adapts, mines, vendors, clean-room-reimplements, sidecar-wraps, or references a public repository -- the acceptance gate that sits above the license\u002Fsecurity boundary, not a substitute for it.\n---\n\n# CoderOS Source Reuse Acceptance Auditor\n\nThis skill defines your role whenever a task involves preparing or\nreviewing a PR that claims to copy, adapt, mine, vendor, clean-room, or\nreference a public repository. It is the acceptance gate a PR must pass\n*after* `docs\u002FSOURCE_COPY_LICENSE_BOUNDARY.md` has already determined\nwhich `ReuseMode` is allowed -- this skill does not re-derive that\nboundary, it audits whether a specific PR actually satisfied it.\n\n## 1. Role boundaries\n\n- **Read `docs\u002FSOURCE_REUSE_ACCEPTANCE_AUDITOR.md` before writing or\n  reviewing a source-reuse PR.** It is the operational rubric; do not\n  substitute general judgment about \"this seems reasonable\" for the\n  explicit rows in its §2.\n- **This skill is downstream of, not a replacement for,\n  `.agents\u002Fskills\u002Fcoderos-source-copy-reviewer\u002FSKILL.md`.** That skill\n  governs mode selection under the license\u002Fsecurity boundary. This skill\n  governs whether a PR that already selected a mode actually completed\n  everything the acceptance rubric requires for that mode.\n- **Every rejection must cite an exact code** from\n  `docs\u002FSOURCE_REUSE_REJECTION_REASONS.md` -- never a paraphrase. This\n  keeps rejections consistent and searchable regardless of which agent\n  lane (Antigravity, Grok, GPT, Codex, Jules, Claude) is reviewing.\n- **An incomplete applicable checklist item is a rejection, not a\n  follow-up TODO.** Do not accept a source-reuse PR \"with a note to\n  finish attribution later.\"\n\n## 2. Consulting the casebook before starting fresh research\n\nBefore researching a candidate system from scratch, check whether\n`docs\u002FCODEROS_HARNESS_PATTERN_CASEBOOK.md` and\n`docs\u002FCODEROS_SOURCE_REUSE_DECISION_CORPUS.md` already cover it -- both\nare living catalogs, not closed sets, and re-deriving a license\u002Fsecurity\nassessment already recorded there wastes effort the corpus exists to\nsave. If the system is covered, cite the existing record id in the PR\nchecklist's \"Casebook\u002Fdecision-corpus record cited\" field rather than\nre-researching it. If extending either corpus with new records, follow\n`docs\u002FCODEROS_AI_CODING_SYSTEMS_REUSE_PROTOCOL.md`'s verification\ndiscipline from your own first research step -- do not add a record on\nthe strength of a search-engine summary alone; fetch the actual primary\nsource, per that protocol §3's account of a hallucinated summary observed\nduring the corpus's own construction. Every new record's\n`source_url`\u002F`license_class` must be added to\n`projects\u002Fcoderos\u002Fsource_mining\u002Fsource_reuse_verified_source_index_v1.json`,\nand the extension itself must pass\n`docs\u002FSOURCE_REUSE_ANTI_FABRICATION_PROTOCOL.md` and\n`docs\u002FSOURCE_REUSE_ANTI_PADDING_PROTOCOL.md` -- if genuinely verifiable\nmaterial runs out before a requested target is reached, report the gap\nexplicitly (`docs\u002FSOURCE_REUSE_VALIDATED_USEFUL_ADDITIONS.md` §6) rather\nthan padding to it.\n\n## 3. Preparing a source-reuse PR (the author's side)\n\n1. Confirm the `ReuseMode` was selected correctly per\n   `docs\u002FSOURCE_COPY_LICENSE_BOUNDARY.md` §3-5 -- do not proceed to this\n   skill's checklist on an unverified mode selection.\n2. Complete every applicable row of\n   `docs\u002FSOURCE_REUSE_ACCEPTANCE_AUDITOR.md` §2, using §3 (minimum\n   attribution fields), §4 (generated-code laundering audit), and §5\n   (file vs. algorithm vs. pattern distinction) as needed to resolve\n   ambiguous cases.\n3. Copy `docs\u002FSOURCE_REUSE_PR_CHECKLIST.md`'s checklist block into the PR\n   description, verbatim, with every box checked or marked `N\u002FA -- reason`.\n4. If any applicable box cannot be checked, do not open the PR as ready\n   for review -- resolve it first, or open as a draft stating exactly\n   what remains outstanding.\n\n## 4. Reviewing a source-reuse PR (the reviewer's side)\n\n1. Verify the PR description contains the completed\n   `docs\u002FSOURCE_REUSE_PR_CHECKLIST.md` block -- a source-reuse PR without\n   it fails `MODE_UNDISCLOSED` at minimum.\n2. Walk every checked box against the actual diff -- do not accept a\n   checked box at face value for attribution completeness (R9-R12),\n   sidecar isolation (R13), or clean-room evidence (R15-R16); verify the\n   artifact is actually present in the diff.\n3. Run the generated-code laundering audit (§4.2 of the auditor doc)\n   yourself for any `copy_pattern` or `clean_room_reimplement` claim --\n   compare structure against the named original source, not just the\n   PR's self-report.\n4. If any applicable rubric row fails, reject citing the exact code from\n   `docs\u002FSOURCE_REUSE_REJECTION_REASONS.md` and the specific factual\n   basis (see that document §8 for the citation format).\n5. Confirm gate composition: this PR still needs to satisfy\n   `docs\u002FPR_EVIDENCE_GATE.md`'s existing Security PR Critic and\n   `test_output` requirements (for runtime changes) -- this auditor adds\n   requirements, it does not remove the pre-existing gate's own.\n6. If the PR is part of a planned CEO Loop \u002F Execution Harness pipeline,\n   confirm the completed checklist and its referenced attribution\u002Fplan\n   artifact are captured as evidence per\n   `docs\u002FSOURCE_REUSE_ACCEPTANCE_AUDITOR.md` §7 -- not left only in the\n   PR description with no durable pointer.\n\n## 5. Lane contribution floors (preparing and reviewing)\n\nAntigravity and Grok have repeatedly shipped PRs that claimed broad\nmining passes while the diff added only a token handful of fixtures --\n`docs\u002FSOURCE_REUSE_LANE_CONTRIBUTION_FLOORS.md` exists specifically to\nstop that pattern. This applies on top of, not instead of, §2-3 above:\n\n- **Before preparing a completion report or marking a PR ready for\n  review**, run `docs\u002FSOURCE_REUSE_LANE_CONTRIBUTION_FLOORS.md` §6's\n  PowerShell snippets and compare the actual additions count against your\n  lane's floor (§2 of that document: Antigravity >4000, Grok >3500, Jules\n  >2000 unless explicitly repair-only, GPT >7000 for a backlog\u002Fdossier,\n  Codex >2500 with targeted tests). **If the floor is not cleared, do not\n  produce a completion report** -- report the work as in-progress,\n  regardless of how clean the diff or how green the tests are.\n- **As a reviewer**, do not accept a PR's own claim of mining breadth at\n  face value. Count the fixture, catalog, and attribution artifacts\n  actually in the diff and compare against what the PR or its checklist\n  header claims was mined -- a wide mismatch is\n  `SHALLOW_FIXTURE_LAUNDERING`. Read the test cases against the\n  originating task's stated scope, not just whether they pass -- tests\n  that only exercise the easiest case when the task asked for\n  license\u002Fsecurity-tier or edge-case coverage is `GREEN_TESTS_MINIMALISM`.\n- **Neither check is satisfied by a large but disproportionate diff.** A\n  PR that clears the line-count floor with padding, boilerplate, or\n  duplicated fixture stubs (see the floors document §5) has not actually\n  met the bar -- verify real corresponding logic\u002Fcoverage, not just the\n  number.\n- **Jules's repair-only exception and Codex's targeted-test requirement**\n  (floors doc §2.1-2.2) are specific enough to check directly: for Jules,\n  quote the originating task's own repair-scoped language before granting\n  the exception; for Codex, confirm the tests run were scoped to the\n  changed surface, not an unscoped full-suite run, unless this is\n  explicitly the final-release integration PR.\n\n## 6. Cross-agent mining reports\n\nIf reviewing or continuing a mining pass reported by a different agent\nlane (Antigravity, Grok, GPT, Codex, Jules), hold it to exactly the same\nreporting bar defined in `docs\u002FSOURCE_REUSE_ACCEPTANCE_AUDITOR.md` §8 --\ndo not accept \"already reviewed elsewhere\" without the required fields\n(source URL, pinned commit SHA, license ID, security tier, selected mode,\noutcome). A `reject`\u002F`defer` outcome from another lane is still a valid,\nreusable record if it has these fields; treat a report missing them as\nnot yet performed, regardless of which lane claims to have done it.\n\n## 6.1 Verifying which \"OpenHands\"\u002F\"OpenBot\"\u002F\"Modular\" you mean\n\nLarge source-mining passes are prone to a specific failure this skill's\nown casebook now has a worked example for: a README's prose describes\ncapabilities that actually live in a *different* repository than the one\nbeing cited, or a common project name collides across multiple unrelated\nGitHub orgs. `docs\u002FCODEROS_HARNESS_SOURCE_REUSE_DOCTRINE.md` §3 and §7\nrecord two real instances (OpenHands's actual runtime split across\n`OpenHands\u002Fopenhands-aci` and `OpenHands\u002Fsoftware-agent-sdk`, distinct\nfrom the frontend-only `OpenHands\u002FOpenHands`; and three unrelated\nprojects all named \"OpenBot\"). Before citing a source's README content\nas authoritative for a specific repository, confirm the directory\nstructure or org-level repo listing actually backs that README's claims\nfor *that* repository, not a same-named sibling.\n\n**Before mining a source by name, check whether that name already means\nsomething inside CoderOS itself.** \"OmniRoute\" turned out to be\nprimarily CoderOS's own internal terminology (`DEV_OMNIROUTE_*` env vars\nin `README.md`) with a real, independently-existing external repository\n(`diegosouzapw\u002FOmniRoute`) also coincidentally sharing the name, plus 19\nfurther unrelated same-named decoy repos. A `Grep` for the candidate name\nacross `docs\u002F`, `repo_dev_runtime\u002F`, and `README.md` costs one call and\nwould catch this class of collision immediately. See\n`docs\u002FCODEROS_EXTERNAL_SOURCE_IDENTITY_PROTOCOL.md` for the full,\nstanding process (identity resolution before license\u002Fsecurity\nclassification) and `docs\u002FCODEROS_HARNESS_SOURCE_REGISTRY_V2.md` for the\nworked OmniRoute\u002FCursor\u002FSlack examples, including Cursor's 8-repo org\nwhere 6 repos are unlicensed and 2 are independently permissive --\nverify license status per-repository, never assume it from the org.\n\n## 7. Auditing first-party autonomy claims\n\nThe same standards apply to CoderOS's own code as to external sources.\n`docs\u002FCODEROS_AUTONOMY_WIRING_AUDIT.md` found `PADDING_SIGNATURE_DETECTED`\nand `GREEN_TESTS_MINIMALISM` -- both defined here for external corpora --\noccurring in first-party code.\n\nWhen a PR or document claims a CoderOS capability operates unattended:\n\n- **Require a named call site and a named failing-if-false test** before\n  accepting the word \"enforced\", \"gates\", \"blocks\" or \"refuses\", per\n  `docs\u002FCODEROS_AUTONOMY_ENFORCEMENT_DOCTRINE.md` §2. Logic that exists\n  but has no caller is \"implemented but unwired\", not enforced.\n- **Do not accept green CI as evidence.** Six of seven autonomy test\n  files in this repository assert nothing while reporting passes.\n- **Do not accept a merged PR as validation.** PR #199 merged carrying\n  four documents padded to exactly 5002 lines each.\n- **Do not accept a commit message as evidence of direction.** PR #205\n  was titled as wiring the autonomy framework and deleted ~1,800 lines of\n  implementation, which PR #206 had to restore.\n- **Check that any doctrine component name resolves to a module path.**\n  \"Merge Queue\" appeared in v1's doctrine with zero occurrences in the\n  codebase; it now exists (`merge_queue.py`, merged via PR #208) --\n  re-verify against the current tree rather than citing v1's finding as\n  still true. See `docs\u002FCODEROS_ENFORCED_AUTONOMY_SPINE_AUDIT_V2.md` for\n  the re-run and `docs\u002FCODEROS_FALSE_AUTONOMY_REGRESSION_TESTS.md` for\n  the per-finding regression status.\n- **A gate that is called every time with a hardcoded input that always\n  passes is worse than a gate that is never called.** v2 found\n  `autonomous_queue_runner_cli.py:16-18` calling its real gate function\n  with `health_green=True, active_leases=[]` literals -- the branching\n  logic runs and looks like enforcement while guaranteeing the pass.\n  Check not just whether a gate is called, but whether its inputs are\n  ever anything other than a hardcoded constant.\n- **Before accepting an in-flight PR's new module, check it does not\n  collide with an already-merged path.** PR #210 independently\n  re-implemented `evidence_trajectory_gate.py` with an incompatible API\n  to the version PR #208 already merged -- `docs\u002FCODEROS_ENFORCED_AUTONOMY_SPINE_AUDIT_V2.md`\n  §2 has the full comparison.\n\nCite findings using the codes in `docs\u002FSOURCE_REUSE_REJECTION_REASONS.md`\n§6.5, which §6.6 confirms apply to first-party PRs on the same terms.\n\n## 8. \"Exceptional human input only\" triggers\n\nHalt and escalate to the operator if:\n\n- The PR's selected mode conflicts with a stop condition in\n  `docs\u002FSOURCE_REUSE_ACCEPTANCE_AUDITOR.md` §6 (copyleft, unknown\n  license, critical\u002Fhigh security risk, offensive tooling) and the\n  operator's request implies proceeding anyway.\n- A generated-code laundering signal (§4 of the auditor doc) is found and\n  the PR author has not obtained explicit operator acceptance of the\n  risk.\n- A clean-room claim shows evidence of context contamination (the\n  original source was open during \"independent\" implementation) per\n  `docs\u002FSOURCE_CLEAN_ROOM_REIMPLEMENTATION_PROTOCOL.md` §4.\n- A different agent lane's mining report is being relied upon without the\n  required fields from §8 of the auditor doc.\n- A lane's contribution floor is unmet and the operator's request implies\n  marking the PR ready for review anyway, or a `SHALLOW_FIXTURE_LAUNDERING`\n  \u002F `GREEN_TESTS_MINIMALISM` finding is disputed by the PR author -- do not\n  resolve the dispute unilaterally in either direction; escalate with the\n  specific counts from `docs\u002FSOURCE_REUSE_LANE_CONTRIBUTION_FLOORS.md` §6.\n- A task's own instructions specify a hard volume\u002Fcount gate (e.g. \"do not\n  report complete below N additions\") for a source-reuse artifact in a way\n  that would require padding or unverified content to satisfy, as\n  distinct from a genuinely achievable, verified scope. Surface the\n  conflict and propose the honestly-scoped alternative rather than\n  silently padding or silently under-delivering without explanation --\n  `docs\u002FCODEROS_HARNESS_PATTERN_CASEBOOK.md` §2 and\n  `docs\u002FCODEROS_SOURCE_REUSE_DECISION_CORPUS.md` §2 are the worked\n  precedent for how this was resolved once already, and\n  `docs\u002FSOURCE_REUSE_ANTI_PADDING_PROTOCOL.md`,\n  `docs\u002FSOURCE_REUSE_ANTI_FABRICATION_PROTOCOL.md`, and\n  `docs\u002FSOURCE_REUSE_VALIDATED_USEFUL_ADDITIONS.md` are the standing\n  protocols this pattern's own review discipline now lives in\n  permanently, rather than only as a one-off precedent.\n","# CoderOS Source Reuse Acceptance Auditor\n\nThis skill defines your role whenever a task involves preparing or\nreviewing a PR that claims to copy, adapt, mine, vendor, clean-room, or\nreference a public repository. It is the acceptance gate a PR must pass\n*after* `docs\u002FSOURCE_COPY_LICENSE_BOUNDARY.md` has already determined\nwhich `ReuseMode` is allowed -- this skill does not re-derive that\nboundary, it audits whether a specific PR actually satisfied it.\n\n## 1. Role boundaries\n\n- **Read `docs\u002FSOURCE_REUSE_ACCEPTANCE_AUDITOR.md` before writing or\n  reviewing a source-reuse PR.** It is the operational rubric; do not\n  substitute general judgment about \"this seems reasonable\" for the\n  explicit rows in its §2.\n- **This skill is downstream of, not a replacement for,\n  `.agents\u002Fskills\u002Fcoderos-source-copy-reviewer\u002FSKILL.md`.** That skill\n  governs mode selection under the license\u002Fsecurity boundary. This skill\n  governs whether a PR that already selected a mode actually completed\n  everything the acceptance rubric requires for that mode.\n- **Every rejection must cite an exact code** from\n  `docs\u002FSOURCE_REUSE_REJECTION_REASONS.md` -- never a paraphrase. This\n  keeps rejections consistent and searchable regardless of which agent\n  lane (Antigravity, Grok, GPT, Codex, Jules, Claude) is reviewing.\n- **An incomplete applicable checklist item is a rejection, not a\n  follow-up TODO.** Do not accept a source-reuse PR \"with a note to\n  finish attribution later.\"\n\n## 2. Consulting the casebook before starting fresh research\n\nBefore researching a candidate system from scratch, check whether\n`docs\u002FCODEROS_HARNESS_PATTERN_CASEBOOK.md` and\n`docs\u002FCODEROS_SOURCE_REUSE_DECISION_CORPUS.md` already cover it -- both\nare living catalogs, not closed sets, and re-deriving a license\u002Fsecurity\nassessment already recorded there wastes effort the corpus exists to\nsave. If the system is covered, cite the existing record id in the PR\nchecklist's \"Casebook\u002Fdecision-corpus record cited\" field rather than\nre-researching it. If extending either corpus with new records, follow\n`docs\u002FCODEROS_AI_CODING_SYSTEMS_REUSE_PROTOCOL.md`'s verification\ndiscipline from your own first research step -- do not add a record on\nthe strength of a search-engine summary alone; fetch the actual primary\nsource, per that protocol §3's account of a hallucinated summary observed\nduring the corpus's own construction. Every new record's\n`source_url`\u002F`license_class` must be added to\n`projects\u002Fcoderos\u002Fsource_mining\u002Fsource_reuse_verified_source_index_v1.json`,\nand the extension itself must pass\n`docs\u002FSOURCE_REUSE_ANTI_FABRICATION_PROTOCOL.md` and\n`docs\u002FSOURCE_REUSE_ANTI_PADDING_PROTOCOL.md` -- if genuinely verifiable\nmaterial runs out before a requested target is reached, report the gap\nexplicitly (`docs\u002FSOURCE_REUSE_VALIDATED_USEFUL_ADDITIONS.md` §6) rather\nthan padding to it.\n\n## 3. Preparing a source-reuse PR (the author's side)\n\n1. Confirm the `ReuseMode` was selected correctly per\n   `docs\u002FSOURCE_COPY_LICENSE_BOUNDARY.md` §3-5 -- do not proceed to this\n   skill's checklist on an unverified mode selection.\n2. Complete every applicable row of\n   `docs\u002FSOURCE_REUSE_ACCEPTANCE_AUDITOR.md` §2, using §3 (minimum\n   attribution fields), §4 (generated-code laundering audit), and §5\n   (file vs. algorithm vs. pattern distinction) as needed to resolve\n   ambiguous cases.\n3. Copy `docs\u002FSOURCE_REUSE_PR_CHECKLIST.md`'s checklist block into the PR\n   description, verbatim, with every box checked or marked `N\u002FA -- reason`.\n4. If any applicable box cannot be checked, do not open the PR as ready\n   for review -- resolve it first, or open as a draft stating exactly\n   what remains outstanding.\n\n## 4. Reviewing a source-reuse PR (the reviewer's side)\n\n1. Verify the PR description contains the completed\n   `docs\u002FSOURCE_REUSE_PR_CHECKLIST.md` block -- a source-reuse PR without\n   it fails `MODE_UNDISCLOSED` at minimum.\n2. Walk every checked box against the actual diff -- do not accept a\n   checked box at face value for attribution completeness (R9-R12),\n   sidecar isolation (R13), or clean-room evidence (R15-R16); verify the\n   artifact is actually present in the diff.\n3. Run the generated-code laundering audit (§4.2 of the auditor doc)\n   yourself for any `copy_pattern` or `clean_room_reimplement` claim --\n   compare structure against the named original source, not just the\n   PR's self-report.\n4. If any applicable rubric row fails, reject citing the exact code from\n   `docs\u002FSOURCE_REUSE_REJECTION_REASONS.md` and the specific factual\n   basis (see that document §8 for the citation format).\n5. Confirm gate composition: this PR still needs to satisfy\n   `docs\u002FPR_EVIDENCE_GATE.md`'s existing Security PR Critic and\n   `test_output` requirements (for runtime changes) -- this auditor adds\n   requirements, it does not remove the pre-existing gate's own.\n6. If the PR is part of a planned CEO Loop \u002F Execution Harness pipeline,\n   confirm the completed checklist and its referenced attribution\u002Fplan\n   artifact are captured as evidence per\n   `docs\u002FSOURCE_REUSE_ACCEPTANCE_AUDITOR.md` §7 -- not left only in the\n   PR description with no durable pointer.\n\n## 5. Lane contribution floors (preparing and reviewing)\n\nAntigravity and Grok have repeatedly shipped PRs that claimed broad\nmining passes while the diff added only a token handful of fixtures --\n`docs\u002FSOURCE_REUSE_LANE_CONTRIBUTION_FLOORS.md` exists specifically to\nstop that pattern. This applies on top of, not instead of, §2-3 above:\n\n- **Before preparing a completion report or marking a PR ready for\n  review**, run `docs\u002FSOURCE_REUSE_LANE_CONTRIBUTION_FLOORS.md` §6's\n  PowerShell snippets and compare the actual additions count against your\n  lane's floor (§2 of that document: Antigravity >4000, Grok >3500, Jules\n  >2000 unless explicitly repair-only, GPT >7000 for a backlog\u002Fdossier,\n  Codex >2500 with targeted tests). **If the floor is not cleared, do not\n  produce a completion report** -- report the work as in-progress,\n  regardless of how clean the diff or how green the tests are.\n- **As a reviewer**, do not accept a PR's own claim of mining breadth at\n  face value. Count the fixture, catalog, and attribution artifacts\n  actually in the diff and compare against what the PR or its checklist\n  header claims was mined -- a wide mismatch is\n  `SHALLOW_FIXTURE_LAUNDERING`. Read the test cases against the\n  originating task's stated scope, not just whether they pass -- tests\n  that only exercise the easiest case when the task asked for\n  license\u002Fsecurity-tier or edge-case coverage is `GREEN_TESTS_MINIMALISM`.\n- **Neither check is satisfied by a large but disproportionate diff.** A\n  PR that clears the line-count floor with padding, boilerplate, or\n  duplicated fixture stubs (see the floors document §5) has not actually\n  met the bar -- verify real corresponding logic\u002Fcoverage, not just the\n  number.\n- **Jules's repair-only exception and Codex's targeted-test requirement**\n  (floors doc §2.1-2.2) are specific enough to check directly: for Jules,\n  quote the originating task's own repair-scoped language before granting\n  the exception; for Codex, confirm the tests run were scoped to the\n  changed surface, not an unscoped full-suite run, unless this is\n  explicitly the final-release integration PR.\n\n## 6. Cross-agent mining reports\n\nIf reviewing or continuing a mining pass reported by a different agent\nlane (Antigravity, Grok, GPT, Codex, Jules), hold it to exactly the same\nreporting bar defined in `docs\u002FSOURCE_REUSE_ACCEPTANCE_AUDITOR.md` §8 --\ndo not accept \"already reviewed elsewhere\" without the required fields\n(source URL, pinned commit SHA, license ID, security tier, selected mode,\noutcome). A `reject`\u002F`defer` outcome from another lane is still a valid,\nreusable record if it has these fields; treat a report missing them as\nnot yet performed, regardless of which lane claims to have done it.\n\n## 6.1 Verifying which \"OpenHands\"\u002F\"OpenBot\"\u002F\"Modular\" you mean\n\nLarge source-mining passes are prone to a specific failure this skill's\nown casebook now has a worked example for: a README's prose describes\ncapabilities that actually live in a *different* repository than the one\nbeing cited, or a common project name collides across multiple unrelated\nGitHub orgs. `docs\u002FCODEROS_HARNESS_SOURCE_REUSE_DOCTRINE.md` §3 and §7\nrecord two real instances (OpenHands's actual runtime split across\n`OpenHands\u002Fopenhands-aci` and `OpenHands\u002Fsoftware-agent-sdk`, distinct\nfrom the frontend-only `OpenHands\u002FOpenHands`; and three unrelated\nprojects all named \"OpenBot\"). Before citing a source's README content\nas authoritative for a specific repository, confirm the directory\nstructure or org-level repo listing actually backs that README's claims\nfor *that* repository, not a same-named sibling.\n\n**Before mining a source by name, check whether that name already means\nsomething inside CoderOS itself.** \"OmniRoute\" turned out to be\nprimarily CoderOS's own internal terminology (`DEV_OMNIROUTE_*` env vars\nin `README.md`) with a real, independently-existing external repository\n(`diegosouzapw\u002FOmniRoute`) also coincidentally sharing the name, plus 19\nfurther unrelated same-named decoy repos. A `Grep` for the candidate name\nacross `docs\u002F`, `repo_dev_runtime\u002F`, and `README.md` costs one call and\nwould catch this class of collision immediately. See\n`docs\u002FCODEROS_EXTERNAL_SOURCE_IDENTITY_PROTOCOL.md` for the full,\nstanding process (identity resolution before license\u002Fsecurity\nclassification) and `docs\u002FCODEROS_HARNESS_SOURCE_REGISTRY_V2.md` for the\nworked OmniRoute\u002FCursor\u002FSlack examples, including Cursor's 8-repo org\nwhere 6 repos are unlicensed and 2 are independently permissive --\nverify license status per-repository, never assume it from the org.\n\n## 7. Auditing first-party autonomy claims\n\nThe same standards apply to CoderOS's own code as to external sources.\n`docs\u002FCODEROS_AUTONOMY_WIRING_AUDIT.md` found `PADDING_SIGNATURE_DETECTED`\nand `GREEN_TESTS_MINIMALISM` -- both defined here for external corpora --\noccurring in first-party code.\n\nWhen a PR or document claims a CoderOS capability operates unattended:\n\n- **Require a named call site and a named failing-if-false test** before\n  accepting the word \"enforced\", \"gates\", \"blocks\" or \"refuses\", per\n  `docs\u002FCODEROS_AUTONOMY_ENFORCEMENT_DOCTRINE.md` §2. Logic that exists\n  but has no caller is \"implemented but unwired\", not enforced.\n- **Do not accept green CI as evidence.** Six of seven autonomy test\n  files in this repository assert nothing while reporting passes.\n- **Do not accept a merged PR as validation.** PR #199 merged carrying\n  four documents padded to exactly 5002 lines each.\n- **Do not accept a commit message as evidence of direction.** PR #205\n  was titled as wiring the autonomy framework and deleted ~1,800 lines of\n  implementation, which PR #206 had to restore.\n- **Check that any doctrine component name resolves to a module path.**\n  \"Merge Queue\" appeared in v1's doctrine with zero occurrences in the\n  codebase; it now exists (`merge_queue.py`, merged via PR #208) --\n  re-verify against the current tree rather than citing v1's finding as\n  still true. See `docs\u002FCODEROS_ENFORCED_AUTONOMY_SPINE_AUDIT_V2.md` for\n  the re-run and `docs\u002FCODEROS_FALSE_AUTONOMY_REGRESSION_TESTS.md` for\n  the per-finding regression status.\n- **A gate that is called every time with a hardcoded input that always\n  passes is worse than a gate that is never called.** v2 found\n  `autonomous_queue_runner_cli.py:16-18` calling its real gate function\n  with `health_green=True, active_leases=[]` literals -- the branching\n  logic runs and looks like enforcement while guaranteeing the pass.\n  Check not just whether a gate is called, but whether its inputs are\n  ever anything other than a hardcoded constant.\n- **Before accepting an in-flight PR's new module, check it does not\n  collide with an already-merged path.** PR #210 independently\n  re-implemented `evidence_trajectory_gate.py` with an incompatible API\n  to the version PR #208 already merged -- `docs\u002FCODEROS_ENFORCED_AUTONOMY_SPINE_AUDIT_V2.md`\n  §2 has the full comparison.\n\nCite findings using the codes in `docs\u002FSOURCE_REUSE_REJECTION_REASONS.md`\n§6.5, which §6.6 confirms apply to first-party PRs on the same terms.\n\n## 8. \"Exceptional human input only\" triggers\n\nHalt and escalate to the operator if:\n\n- The PR's selected mode conflicts with a stop condition in\n  `docs\u002FSOURCE_REUSE_ACCEPTANCE_AUDITOR.md` §6 (copyleft, unknown\n  license, critical\u002Fhigh security risk, offensive tooling) and the\n  operator's request implies proceeding anyway.\n- A generated-code laundering signal (§4 of the auditor doc) is found and\n  the PR author has not obtained explicit operator acceptance of the\n  risk.\n- A clean-room claim shows evidence of context contamination (the\n  original source was open during \"independent\" implementation) per\n  `docs\u002FSOURCE_CLEAN_ROOM_REIMPLEMENTATION_PROTOCOL.md` §4.\n- A different agent lane's mining report is being relied upon without the\n  required fields from §8 of the auditor doc.\n- A lane's contribution floor is unmet and the operator's request implies\n  marking the PR ready for review anyway, or a `SHALLOW_FIXTURE_LAUNDERING`\n  \u002F `GREEN_TESTS_MINIMALISM` finding is disputed by the PR author -- do not\n  resolve the dispute unilaterally in either direction; escalate with the\n  specific counts from `docs\u002FSOURCE_REUSE_LANE_CONTRIBUTION_FLOORS.md` §6.\n- A task's own instructions specify a hard volume\u002Fcount gate (e.g. \"do not\n  report complete below N additions\") for a source-reuse artifact in a way\n  that would require padding or unverified content to satisfy, as\n  distinct from a genuinely achievable, verified scope. Surface the\n  conflict and propose the honestly-scoped alternative rather than\n  silently padding or silently under-delivering without explanation --\n  `docs\u002FCODEROS_HARNESS_PATTERN_CASEBOOK.md` §2 and\n  `docs\u002FCODEROS_SOURCE_REUSE_DECISION_CORPUS.md` §2 are the worked\n  precedent for how this was resolved once already, and\n  `docs\u002FSOURCE_REUSE_ANTI_PADDING_PROTOCOL.md`,\n  `docs\u002FSOURCE_REUSE_ANTI_FABRICATION_PROTOCOL.md`, and\n  `docs\u002FSOURCE_REUSE_VALIDATED_USEFUL_ADDITIONS.md` are the standing\n  protocols this pattern's own review discipline now lives in\n  permanently, rather than only as a one-off precedent.\n",{"name":5,"description":7},"main","b980e90b49ea7c0639094f3060ced5aaf772a571",".agents\u002Fskills\u002Fcoderos-source-reuse-auditor",0,"https:\u002F\u002Fgithub.com\u002FChristianV997\u002FCoderOS","https:\u002F\u002Fgithub.com\u002FChristianV997","2026-10-07 15:44:03",{"source":8,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add ChristianV997\u002FCoderOS --skill coderos-source-reuse-auditor",{"tool":36,"command":37},"bunx","bunx skills add ChristianV997\u002FCoderOS --skill coderos-source-reuse-auditor",{"tool":39,"command":40},"pnpm","pnpm dlx skills add ChristianV997\u002FCoderOS --skill coderos-source-reuse-auditor","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FChristianV997\u002FCoderOS\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcoderos-source-reuse-auditor\u002FSKILL.md",true,[],[45,52,59,66,73,80],{"skillNo":46,"name":47,"nameZh":6,"description":48,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":49,"categories":50,"keywords":51,"publishedAt":18},"01M7A0MH3MZXKX7B9AWKFNM5DB","coderos","Workspace-native skill for operating, extending, and safe parallel implementation on CoderOS itself. Use when selecting non-overlapping implementation slices, navigating CoderOS governance, scout, work-order, sandbox, and skill-pack seams, and executing safe changes.","https:\u002F\u002Fgithub.com\u002FChristianV997\u002FCoderOS\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcoderos\u002FSKILL.md",[],[],{"skillNo":53,"name":54,"nameZh":6,"description":55,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":56,"categories":57,"keywords":58,"publishedAt":18},"01M7A0ME1KR3V3M79FXR0FHM9Q","coderos-workspace-operator","Skill for interacting with the local CoderOS Operator Workspace to coordinate tasks and prevent collisions.","https:\u002F\u002Fgithub.com\u002FChristianV997\u002FCoderOS\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcoderos-workspace-operator\u002FSKILL.md",[],[],{"skillNo":60,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":18},"01M7A0M9GV17YZCTWS23Y9SY1K","coderos-workorder","Workspace-native skill for drafting valid typed CoderOS WorkOrder JSON payloads and exercising dry-run work-order dispatch safely without live execution or runtime mutation.","https:\u002F\u002Fgithub.com\u002FChristianV997\u002FCoderOS\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcoderos-workorder\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":18},"01M7A0M6H61K6K6C8FKKVP0F29","coderos-worker-adapters","Generates deterministic safe command packets for various workers","https:\u002F\u002Fgithub.com\u002FChristianV997\u002FCoderOS\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcoderos-worker-adapters\u002FSKILL.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":18},"01M7A0M4DY9PCD2D0NTH2Q8WC0","coderos-work-order-factory","Draft and validate deterministic CoderOS work-order factory artifacts without dispatch or provider activation.","https:\u002F\u002Fgithub.com\u002FChristianV997\u002FCoderOS\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcoderos-work-order-factory\u002FSKILL.md",[],[],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":18},"01M7A0KTEKMAH8NFAT80MZ2KTG","coderos-source-mining-wave4","Workspace-native skill for operating, auditing, and generating Source-Mining Wave 4 catalog records, retrofit decisions, and implementation waves for external agent OS systems.","https:\u002F\u002Fgithub.com\u002FChristianV997\u002FCoderOS\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcoderos-source-mining-wave4\u002FSKILL.md",[],[]]