A2 — Spec Review
Skill2 starsChanged 5 months agoApache-2.0View on GitHub
Mechanical FR/AC compliance. Walk every functional requirement and acceptance criterion and render a binary Pass/Fail with evidence. No judgment about quality, direction, or architecture.
In this repository22 other files it gives its agents.
- A1 — Design & Master-Spec ConsistencySkill
- A3 — PolishSkill
- A4 — North Star DirectionSkill
- A5 — Web AuditSkill
- AcceptSkill
- AmendSkill
- BrainstormSkill
- ClarifySkill
- ClosureSkill
- ConstituteSkill
- DesignSkill
- DiscoverSkill
- FeatureSkill
- Final — The F-PassSkill
- HelpSkill
- ImplementSkill
- InitSkill
- MicroSkill
- North StarSkill
- ResumeSkill
- SpecifySkill
- StatusSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 22
Write the report to audits/{scope}-a2.md: a coverage table listing every FR/AC with Pass/Fail and its evidence, a finding per failure with …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
A2 — Spec Review (FR/AC Compliance)
The mechanical compliance check. A2 walks every Functional Requirement (FR) and every Acceptance Criterion (AC) in the master spec and renders a binary verdict — Pass or Fail — against the implementation, with evidence. It makes no judgment about code quality (that is A3), direction (A4), or architecture (A1). Exhaustive and non-judgmental: each requirement either is satisfied by the code or it is not.
Inputs
spec.md(required — the source of all FRs/ACs) · all micro specs in scope · the actual code files they reference- The constitution · prior audit reports for this scope (e.g. the A1 report) for cumulative context · the project mode
Process
- Build the requirement checklist: extract every FR and AC that applies to this scope (for a group, the FRs/ACs its micro specs cover).
- Scan each file against each requirement — an exhaustive file-by-file pass. Produce a per-requirement verdict (Pass/Fail) with specific evidence.
- Review the scan for accuracy — correct false positives and false negatives, add context. (Two stages: a fast exhaustive pass, then a careful review pass.)
- (Brownfield) same File-Operations-table scoping and Inherited/Introduced classification as A1.
- Write the report to
audits/{scope}-a2.md: a coverage table listing every FR/AC with Pass/Fail and its evidence, a finding per failure with reasoning trace, and an overall verdict.
Discipline
A2 is binary and evidence-bound. A well-formed finding cites (1) the exact FR/AC ID with its wording, (2) the implementation evidence that fails it, and (3) a Pass/Fail verdict with no ambiguity. If a requirement is itself ambiguous, A2 flags the ambiguity rather than guessing intent — it never invents a passing or failing interpretation to resolve unclear spec language.
Shared machinery
Finding format, the zero-tolerance loop (loops to zero findings), and conviction scoring are shared across all lenses — see docs/the-audit-system.md. A2 may run alongside A1, but A3 still waits on A1 being clean.
Natural language triggers
- "a2 audit group 1"
- "spec review"
- "check FR/AC compliance"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.

