agentleFS
komlukkomluk/scaffolding/skills/spec-review/SKILL.md

spec-review

Skill16 starsChanged 6 days agoMITView on GitHub

Verify an implementation against its OpenSpec artifacts. TRIGGER when: checking completed work against design.md and tasks.md. SKIP: security-specific review (use security-review-checklists); executing tasks (use spec-develop).

47 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use get-pr-reviews, push-everything and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • saffron-healthSkill

    get-pr-reviews

    Get nicely formatted output of the last review comments on a PR. Used internally by other skills.

    Also in 3 projects that use this

  • saffron-healthSkill

    push-everything

    Stage ALL changes in the repository (not just session changes), commit, and push. Use when you need to commit all pending changes.

    Also in 3 projects that use this

  • NVIDIASkill

    generate-spec

    Produce a feature specification covering API changes, backward compatibility, cross-language impact, performance, and a file-level change plan. Use when designing a new feature, planning a significant change, or the user asks to spec something out.

    Also in 3 projects that use this

  • telefrekSkill

    spec-init

    Bootstrap the spec corpus for this project

    Also in 3 projects that use this

  • AWSSkill

    implement-spec

    Implement the result of /to-spec and /to-tickets in code.

    Also in 8 projects that use this

  • ArvorCoSkill

    specify

    Create feature specification from natural language description. Use when starting a new feature or creating a spec. Triggers on: create spec, specify feature, new feature spec.

    Also in 3 projects that use this

  • +2 more from saffron-health

The file

We split this skill into the rules it gives an agent

9 rules. 0 are common, 9 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 43 edited versions we read keep these.

  • 9

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When acting on review feedback:

    line 72
  2. Distinctive:

    Disagree with evidence: - if a finding is wrong or conflicts with the spec, reply with the reason (file/line, spec reference) instead of …

    line 75
  3. Distinctive:

    One fix, one check: - after each change re-run the relevant tests.

    line 76
  4. Distinctive:

    Load artifacts: - Read design.md, tasks.md, proposal.md

    line 82
  5. Distinctive:

    Always note which checks were skipped and why.

    line 125
  6. Distinctive:

    Correctness: Use keyword search + file path analysis.

    line 130
  7. Distinctive:

    Coherence: Look for glaring inconsistencies, don't nitpick style

    line 131
  8. Distinctive:

    False positives: Prefer SUGGESTION over WARNING, WARNING over CRITICAL when uncertain

    line 132
  9. Distinctive:

    Actionability: Every issue must have a specific recommendation with file references

    line 133

“Common” is measured: kept by at least 6 in 10 of the 43 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

OpenSpec Verification

Guide for verifying that implementation matches spec artifacts.

Input Files

FileRequiredPurpose
{specs_path}/design.mdYesRequirements and scenarios to verify
{specs_path}/tasks.mdYesCompletion checklist
{specs_path}/proposal.mdOptionalOriginal intent reference

Path Enforcement: The specs_path MUST be .scaffolding/conversations/{UUID}/specs/ where {UUID} is a valid UUID (format: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx). NEVER use descriptive folder names.

.scaffolding/conversations/ is a symlink to <share>/projects/<repo-name>/conversations (paths below are unchanged; always go through the symlink).

Three Verification Dimensions

1. Completeness

Question: Are all tasks done and all requirements covered?

CheckMethodIssue Level
All checkboxes marked [x]Parse tasks.mdCRITICAL if incomplete
All requirements have codeSearch codebase for keywordsCRITICAL if missing
All new files existVerify file paths from tasksCRITICAL if missing

2. Correctness

Question: Does the code do what the spec says?

CheckMethodIssue Level
GIVEN/WHEN/THEN satisfiedTrace scenario through codeWARNING if divergent
Tests cover scenariosMatch test names to scenariosWARNING if uncovered
Edge cases handledCheck error paths in codeWARNING if missing
Validation commands passRun pytest / npm run validateCRITICAL if failing

3. Coherence

Question: Does the code match design decisions?

CheckMethodIssue Level
Design decisions followedCompare Decisions section to codeWARNING if violated
Patterns consistentCheck naming, structure, styleSUGGESTION
No undocumented changesDiff scope vs design scopeWARNING if extra
No design deviationsCross-reference architectureWARNING if different

Two-Stage Review

Review in two ordered stages. Do not start stage 2 until stage 1 passes.

StageQuestionBasisBlocks on
1. Spec complianceDoes it build exactly what was asked - nothing missing, nothing extra?design.md requirements/scenarios, tasks.mdMissing or divergent requirement, undocumented scope
2. Code qualityIs it well built - readable, tested, secure, maintainable?Project conventions, tests, patternsBugs, missing tests, unsafe code; style nits are SUGGESTION

Why this order: polishing code that implements the wrong thing is wasted work, and quality findings on code that will change after a spec fix are noise. Stage 1 maps to the Completeness + Correctness + Coherence dimensions above; stage 2 is the code-quality pass layered on top. Report stage results separately.

Receiving Code Review (implementer side)

When acting on review feedback:

  1. Read all items first, then address them in order: CRITICAL, WARNING, SUGGESTION.
  2. Verify before changing - confirm each finding against the code; do not apply a suggestion blindly.
  3. Disagree with evidence - if a finding is wrong or conflicts with the spec, reply with the reason (file/line, spec reference) instead of silently ignoring or complying.
  4. One fix, one check - after each change re-run the relevant tests; do not batch unrelated fixes.
  5. Ask when unclear - an ambiguous item gets a question, not a guess.
  6. Report back per item: fixed / disputed (why) / deferred (why). No performative agreement; state what changed.

Verification Process

  1. Load artifacts - Read design.md, tasks.md, proposal.md
  2. Check completeness - Parse checkboxes, search for requirement implementations
  3. Check correctness - Trace each scenario through code, verify test coverage
  4. Check coherence - Compare decisions to implementation, check patterns
  5. Generate report - Summarize findings with issue levels

Report Format

## Verification Report

### Summary
| Dimension    | Status              |
|--------------|---------------------|
| Completeness | X/Y tasks, N reqs   |
| Correctness  | M/N scenarios pass   |
| Coherence    | Followed / N issues  |

### Critical Issues
| # | Dimension | Issue | File | Recommendation |
|---|-----------|-------|------|----------------|
| 1 | Completeness | Task 2.3 incomplete | - | Complete or mark blocked |

### Warnings
| # | Dimension | Issue | File | Recommendation |
|---|-----------|-------|------|----------------|
| 1 | Correctness | Scenario X not tested | test_foo.py | Add test case |

### Suggestions
- [Pattern deviation details with file reference]

### Assessment
[CRITICAL: N issues | WARNINGS: N | Ready for archive: Yes/No]

Graceful Degradation

Available ArtifactsChecks Performed
tasks.md onlyCompleteness (checkboxes) only
tasks.md + design.mdCompleteness + Correctness
All threeAll three dimensions

Always note which checks were skipped and why.

Verification Heuristics

  • Completeness: Focus on objective items (checkboxes, requirement lists)
  • Correctness: Use keyword search + file path analysis; don't require certainty
  • Coherence: Look for glaring inconsistencies, don't nitpick style
  • False positives: Prefer SUGGESTION over WARNING, WARNING over CRITICAL when uncertain
  • Actionability: Every issue must have a specific recommendation with file references

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.