compliance-check
Skill4 starsChanged 3 months agoMITView on GitHub
Diff requirements against draft coverage and update the compliance matrix. Run after proposal-writer, before red-team-review. Reads working/compliance-matrix.md + drafts/; writes reviews/compliance-gaps.md.
9 other projects use a version of this.
Adopted by
Most recently taken up by
- nisarth/claude-skills-storetheir own version, first seen Nov 8, 2026
- EvezArt/evez-skillstheir own version, first seen Nov 1, 2026
- samsam007b/claude-code-standards-kittheir own version, first seen Oct 13, 2026
- microsoft/aibast-agents-librarytheir own version, first seen Oct 13, 2026
- anthropics/knowledge-work-pluginstheir own version, first seen Oct 13, 2026
- me2resh/apexyardtheir own version, first seen Oct 8, 2026
Other versions9 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 9
- anthropics/knowledge-work-plugins★ 36ksame name
- me2resh/apexyard★ 542same name
- alebgl77/claude-inc★ 21same name
- microsoft/aibast-agents-library★ 8same name
- EvezArt/evez-skills★ 3same name
- JustineDevs/premortem★ 2same name
- samsam007b/claude-code-standards-kit★ 2same name
- evgenii-studitskikh/Claude-Code-SaaS-Studio★ 1same name
- nisarth/claude-skills-store★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository30 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- adversarial-reviewSkill
- capture-demand-signalsSkill
- capture-intentSkill
- capture-portal-structureSkill
- capture-scorecardSkill
- capture-submissionSkill
- competitor-assessmentSkill
- customer-intelSkill
- dashboardSkill
- evidence-checkSkill
- export-proposalSkill
- import-from-captureSkill
- narrative-spineSkill
- new-proposalSkill
- opportunity-quick-lookSkill
- past-performanceSkill
- pricing-analystSkill
- proposal-editorSkill
- proposal-graphicsSkill
- proposal-managerSkill
- proposal-patcherSkill
- proposal-solution-architectSkill
- proposal-storyboardSkill
- proposal-writerSkill
- red-team-reviewSkill
- setup-companySkill
- statusSkill
- submission-summarySkill
- technical-reviewSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
17 rules. 0 are common, 7 are distinctive, and 10 are very specific to this project.
0
Common
Most of the 7 edited versions we read keep these.
7
Distinctive
This author's own choices. Worth a look.
10
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 13
Ensure every "shall" from the solicitation has a home in the draft.
- Very specific: line 17
After each proposal-writer pass (checks whether new drafts closed open reqs)
- Very specific: line 18
Before every red-team-review invocation (Pink is mostly this, but formal)
- Very specific: line 19
Before final submission (as a gate — zero Gap rows, all rationales in place)
- Very specific: line 56
Read working/proposal-type.md.
- Very specific: line 59
Read working/compliance-matrix.md.
- Distinctive: line 63
If Section is blank and no draft references the Req ID → Gap
- Distinctive: line 64
If Section is assigned but no draft file exists for that section → Planned
- Distinctive: line 65
If a draft file exists and contains the Req ID (or its verbatim phrase, case-insensitive, first 50 chars) → search for substantive …
- Distinctive: line 66
If Section is assigned but draft mentions the req only in passing → Partial
- Distinctive: line 67
If user previously flipped to Exception, preserve it but require non-empty Evidence column (otherwise flag)
- Very specific: line 69
Write reviews/compliance-gaps.md with gap/partial/exception lists and recommended owner for each (proposal-writer for Gaps, humans for …
- Very specific: line 76
If the req uses "shall describe/provide/demonstrate," the draft has a matching descriptive/demonstrative passage
- Distinctive: line 78
If uncertain, mark Drafted (not Covered) and flag for human review.
- Very specific: line 82
Do not invent coverage. If unsure, mark Partial or Drafted, not Covered.
- Very specific: line 83
Do not silently drop rows.
- Distinctive: line 84
Never rewrite the Requirement verbatim column — that's the source text.
“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
/compliance-check
Purpose
Ensure every "shall" from the solicitation has a home in the draft. Closes the loop between proposal-manager (which seeds the matrix) and proposal-writer (which should be filling it). Produces a gap list that red-team treats as P0 findings.
When to run
- After each
proposal-writerpass (checks whether new drafts closed open reqs) - Before every
red-team-reviewinvocation (Pink is mostly this, but formal) - Before final submission (as a gate — zero
Gaprows, all rationales in place)
What it reads
working/proposal-type.md— forcompliance_sourcesand the skipped checkworking/compliance-matrix.md— the seeded + updated matrixdrafts/*.md— to verify actual coverageworking/proposal-plan.md— for cross-check of eval criteria
What it writes
- Updates
working/compliance-matrix.md:- Recomputes the Status column for every row by scanning
drafts/ - Regenerates the Summary counters block at the bottom
- Updates the
Last updatedfield
- Recomputes the Status column for every row by scanning
- Updates
working/compliance-matrix.json— structured sidecar conforming toreference/schemas/compliance-matrix.schema.json. Written atomically on every run, fully regenerated from the current.mdstate plus the draft scan. Required fields:schema_version: "compliance-matrix.v1"generated_by: "compliance-check"generated_at: ISO-8601 timestampproposal_name,type_id,compliance_sources(fromworking/proposal-type.md)rows— every matrix row withreq_id,source,requirement,section,page,status,evidence, and (Phase C)evidence_refssummary— counters derived from the rows
- Creates/overwrites
reviews/compliance-gaps.mdwith:- Total + counters
- Gap list (requirements with Status =
Gap) - Planned-but-undrafted list (Status =
PlannedAND at least one draft exists) - Partials (Status =
Partial) - Exception log (Status =
Exception— verify rationale present) - Recommendations per finding
- Creates/overwrites
working/compliance-matrix.xlsx— sortable/filterable Excel version for sharing with teammates. Delegate toanthropic-skills:xlsxusing the spec atreference/office-templates/compliance-matrix-template.md. The xlsx reads from the JSON sidecar (not re-parses markdown) for consistency. This runs on every/compliance-checkinvocation, not just on export, so teammates always have a current spreadsheet. (The final submission-package copy atfinal/xlsx/compliance-matrix.xlsxis produced by/export-proposaland may be styled differently per branded template.)
.md, .json, .xlsx all stay in sync
Every /compliance-check run produces all three forms from the same in-memory representation. Humans can hand-edit the .md between runs (to flip a row to Exception, add Evidence, etc.); on the next run, those hand edits are preserved in the regenerated .md and propagated to .json + .xlsx. Do not hand-edit the .json or .xlsx — they're derived.
Algorithm
- Read
working/proposal-type.md. Ifcompliance_sourcesis empty list, exit with:Skipped: proposal type
<type_id>has no compliance sources. No matrix expected. - Read
working/compliance-matrix.md. If it doesn't exist, exit with:No compliance matrix found. Run
/proposal-managerfirst to seed it. - For each row in the matrix:
- If Section is blank and no draft references the Req ID → Gap
- If Section is assigned but no draft file exists for that section → Planned
- If a draft file exists and contains the Req ID (or its verbatim phrase, case-insensitive, first 50 chars) → search for substantive treatment (heading, paragraph ≥50 words that mentions key terms) → Drafted or Covered based on completeness heuristic
- If Section is assigned but draft mentions the req only in passing → Partial
- If user previously flipped to Exception, preserve it but require non-empty Evidence column (otherwise flag)
- Recompute counters. Update the matrix file.
- Write
reviews/compliance-gaps.mdwith gap/partial/exception lists and recommended owner for each (proposal-writer for Gaps, humans for Exceptions).
Coverage heuristic (substantive treatment)
A requirement is Covered (not merely Drafted) when the draft:
- Contains a heading or sub-heading that addresses it, AND
- Has ≥50 words of substantive discussion mentioning the requirement's key nouns/verbs, AND
- If the req uses "shall describe/provide/demonstrate," the draft has a matching descriptive/demonstrative passage
If uncertain, mark Drafted (not Covered) and flag for human review.
Output discipline
- Do not invent coverage. If unsure, mark
PartialorDrafted, notCovered. - Do not silently drop rows. If the matrix has rows from a prior proposal-manager pass that don't appear in current drafts, they stay as
PlannedorGap. - Never rewrite the Requirement verbatim column — that's the source text.
On completion
Append to working/activity.md:
## <timestamp> — compliance-check — <N total>: <covered>C / <drafted>D / <planned>P / <gap>G → reviews/compliance-gaps.mdThen tell the user the headline counts and the #1 gap to close first.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

