Spec review
Review requirements for quality, consistency, and completeness.From its SKILL.md
npx -y skills add agent-ix/quoin --skill spec-reviewAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 20 stars20 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
4.1 KB, as published. Nobody here has run it
Review Requirements
Use this skill to validate requirements artefacts before implementation.
Checklist
See references/checklist.md for the detailed quality gates.
Choose the review set
Before reviewing, present these options to the user and let them choose:
- base — the Checklist only: ID formats, US/FR/TC quality, and the six coverage rules. No analysis skills.
- all —
baseplus all seven analyses below. - subset —
baseplus the analyses the user picks from the seven below.
The seven analyses:
| Analysis | Skill |
|---|---|
| failure-domain | spec-failure-domain-analysis |
| integrity | spec-integrity-analysis |
| dependency | spec-dependency-analysis |
| evidence | spec-evidence-analysis |
| risk-complexity | spec-risk-complexity-analysis |
| scope-boundary | spec-scope-boundary-analysis |
| ears-conformance | spec-ears-analysis |
Process
- Automated Checks (always — part of
base):- Validate ID formats (US/FR/TC/AC/CON).
- Detect duplicates or gaps.
- Validation link integrity.
- Base Review (always):
- Read each artifact.
- Verify against the Checklist.
- Verify Test Coverage (6 Rules).
- Selected Analyses:
- Run each analysis the user chose. Prefer running them in parallel.
- Fetch the template from quoin once with
quoin write --types SpecReview— use its skeleton + schema as the contract (do not invent the format). - Write one
SpecReviewdocument per analysis tospec/reviews/<analysis>.md(plusspec/reviews/base.mdfor the base checklist). Findings go in a validated## Findingstable (| ID | Severity | Summary | Refs |,FND-NNNids, Severity ∈low/medium/high).
- Validate + record:
- Run
quire validate --scope <repo> "spec/**/*.md"and fix any errors.
- Run
--scopeis the repository root, and must be passed explicitly. Since quire-cli v0.16.0 (quire-rs CR-045) the command derives two roots from it and never interchanges them: spec documents are read from<repo>/speconly, trace tags from the source tree at<repo>excludingspec/. A repo with nospec/exits with a diagnostic naming the missing document root rather than scanning the whole tree, and a matrix outsidespec/(a fixture, aplan/copy) mints nothing. A relative glob resolves under--scopeonly in scoped mode (no--module); with--moduleit resolves against the process working directory, and an omitted--scopedefaults to.— so a run launched from a parent directory validates the wrong tree and exits 0 for whatever it matched. Checkquire --version≥ 0.16.0 before relying on any of this; ≤ 0.15.0 has the pre-split traversal semantics.
- The docs sync into filament-core as `SpecReview` artifacts; do not proceed to
implementation until reviewed.
Enforced runs (workflow mode)
quoin review drives this through ix-flow, which records the chosen set and hard-blocks
acceptance until every selected analysis has produced a validated SpecReview doc (the
selected_analyses_covered gate). Use it when you want the choice enforced rather than
trusted to discipline. The per-analysis SpecReview structure + the findings table are
defined in the workflow skill (workflow-assets/skills/review/SKILL.md).
Common Issues
- Vague Criteria: "Fast", "User-friendly". -> Make measurable.
- Missing Errors: Only happy path. -> Add error modes.
- No Coverage: AC without TC. -> Add tests.
- Inconsistent IDs:
US-1instead ofUS-001. -> Fix formatting.
Output contract
The SpecReview this skill emits carries analysis: base — the generic review value — this skill runs the analysis lenses and writes one document per lens, so it is the default rather than a lens of its own.
Stated here because nothing checked it (quoin#202): a skill and its declared output were unlinked, so drift between them was invisible until a document failed validation in an unrelated session.