agentsclimarketplace

Spec app review

Skill agent-ix/quoin/skills/spec-app-review

Review an application-type spec for completeness against UI rendering requirements and golden-path structure.From its SKILL.md

Install
npx -y skills add agent-ix/quoin --skill spec-app-review

Assembled 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

2.7 KB, as published. Nobody here has run it

Review Application Specification

Validate a component_type: application spec against what the spec-editor-ui actually renders and the structural requirements from specify.

When to Use

  • After creating or updating an application spec
  • Before submitting a spec for formal review
  • When debugging why sections appear empty in the spec-editor-ui Application Detail view

Process

  1. Locate spec: Read spec/spec.md in the target repo. Verify component_type: application in frontmatter.
  2. Run checklist: Evaluate every item in references/checklist.md.
  3. Report findings: Group issues by severity (FAIL = will break UI, WARN = degraded display, INFO = quality improvement).
  4. Fix or flag: If instructed to fix, apply corrections. Otherwise, output the report.

Quick Reference

The spec-editor-ui ApplicationDetailPage has 7 tabs. Each tab extracts data from specific sections/artifacts:

TabData Source
OverviewPurpose, In Scope, Out of Scope, Operational Boundaries, API Endpoints, Components (catalog), Standards (frontmatter)
ArchitectureMermaid diagrams (tagged logical/deployment), FR diagrams, Aggregated Objects (catalog)
IntegrationArtifacts with tags: [integration] in frontmatter
SecurityTrust Boundaries table, artifacts with tags: [security], Standards
RequirementsStR/US/FR/NFR tables parsed from spec.md
DecisionsADR artifacts (type: ADR) or inline ADR sections
SpecRaw markdown rendering

Key Parsing Rules (Quire Library)

These are the exact parser behaviors that determine whether content renders:

  • Out of Scope: parseDelegations() only recognizes (Handled by X) or (Delegated to X) patterns. Other parentheticals won't extract delegation targets.
  • In Scope: parseBulletList() default pattern is bold-description (**Title** — Description). Plain bullets fall back to title-only.
  • Operational Boundaries: parseBulletList({ pattern: 'bold-colon' }) requires **Name**: detail format.
  • Trust Boundaries: parseTable() on section "Trust Boundaries" — expects 3 columns: Role, Capabilities, Restrictions.
  • Mermaid tags: findDiagramByTag() matches %% @type: logical and %% @type: deployment comments in mermaid blocks.
  • Integration tab: Filters artifacts by tags: [integration] in YAML frontmatter (case-insensitive).
  • Security tab: Filters artifacts by tags: [security] in YAML frontmatter.

Golden Reference

Use a known well-formed application spec in your org as a comparison reference when reviewing.

Keep looking

Skills are one crate of 611,919. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.