How do I install the Requirement Compliance Review skill?

Run npx skillmds@latest add tuana-vn/requirement-compliance-review in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Requirement Compliance Review safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Requirement Compliance Review?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Requirement Compliance Review free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Requirement Compliance Review?

tuana-vn (@tuana-vn) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Requirement Compliance Review by tuana-vn

Name: Requirement Compliance Review
Author: tuana-vn

Requirement Compliance Review

Compare a user/customer requirement against a patch's reconstructed implemented behavior and the current-system baseline. Use to determine whether development is correct, complete, over-scoped, or likely to regress unrelated behavior.

tuana-vn Updated Sep 17, 2026

File contents

Requirement vs Implementation Compliance

Inputs

requirement/specification
patch/diff
patch impact review if available
current baseline memory

Normalize Requirement

Create atomic rules R-001, R-002, ...

For each capture:

actor/trigger
preconditions
expected rule
expected result
failure/exception behavior
compatibility requirement
ambiguity

Do not invent acceptance criteria.

Requirement Fidelity Gate

For every normalized mandatory field/behavior/condition, point to the exact authoritative requirement clause that requires it. If no clause exists, it must not be represented as a requirement obligation. It may appear only as an explicitly labeled PROPOSED design option/guidance.

Required threshold:

UNSUPPORTED_REQUIREMENT_EXPANSION = 0

Before a ready/pass verdict, reconcile executive summaries and aggregate counts with the row-level matrix. A claim such as "all requirements fail" is invalid when any matrix row is PASS/NO_GAP.

Compare

Use implementation rules from the patch impact review where possible.

Central matrix:

Req ID	Expected Behavior	Implemented Behavior	Status	Gap	Risk	Evidence

Statuses:

PASS
PARTIAL
FAIL
NOT IMPLEMENTED
OVER-IMPLEMENTED
UNKNOWN

Detect

missing conditions
missing branches
wrong defaults
over-broad changes
compatibility regressions
requirement ambiguity
wrong integration/client selection
error-path mismatch
observability/audit mismatch

Overall Verdict

Use:

COMPLIANT
MOSTLY COMPLIANT
PARTIALLY COMPLIANT
NON-COMPLIANT
CANNOT VERIFY

Output

docs/reverse-engineering/reviews/<patch>-requirement-compliance.md

Never update baseline memory from an unmerged patch.

---name: requirement-compliance-reviewdescription: Compare a user/customer requirement against a patch's reconstructed implemented behavior and the current-system baseline. Use to determine whether development is correct, complete, over-scoped, or likely to regress unrelated behavior.---# Requirement vs Implementation Compliance## Inputs- requirement/specification- patch/diff- patch impact review if available- current baseline memory## Normalize RequirementCreate atomic rules `R-001`, `R-002`, ...For each capture:- actor/trigger- preconditions- expected rule- expected result- failure/exception behavior- compatibility requirement- ambiguityDo not invent acceptance criteria.## Requirement Fidelity GateFor every normalized mandatory field/behavior/condition, point to the exact authoritative requirementclause that requires it. If no clause exists, it must not be represented as a requirement obligation.It may appear only as an explicitly labeled `PROPOSED` design option/guidance.Required threshold:```textUNSUPPORTED_REQUIREMENT_EXPANSION = 0```Before a ready/pass verdict, reconcile executive summaries and aggregate counts with the row-levelmatrix. A claim such as "all requirements fail" is invalid when any matrix row is PASS/NO_GAP.## CompareUse implementation rules from the patch impact review where possible.Central matrix:| Req ID | Expected Behavior | Implemented Behavior | Status | Gap | Risk | Evidence ||---|---|---|---|---|---|---|Statuses:- PASS- PARTIAL- FAIL- NOT IMPLEMENTED- OVER-IMPLEMENTED- UNKNOWN## Detect- missing conditions- missing branches- wrong defaults- over-broad changes- compatibility regressions- requirement ambiguity- wrong integration/client selection- error-path mismatch- observability/audit mismatch## Overall VerdictUse:- COMPLIANT- MOSTLY COMPLIANT- PARTIALLY COMPLIANT- NON-COMPLIANT- CANNOT VERIFY## Output`docs/reverse-engineering/reviews/<patch>-requirement-compliance.md`Never update baseline memory from an unmerged patch.

Frequently asked questions