- Home
- Skills
- Dceoy
- Speckit Agent Skills
- Speckit Analyze
speckit-analyze_skill
- Shell
11
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill dceoy/speckit-agent-skills --skill speckit-analyze- SKILL.md7.4 KB
Overview
This skill performs a strict, read-only cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation. It identifies inconsistencies, ambiguities, duplications, coverage gaps, and constitution violations, then produces a compact, actionable analysis report. The skill never edits files and highlights CRITICAL issues that must be resolved before implementation.
How this skill works
The skill initializes by running the prerequisite check script to locate feature docs, then loads minimal required sections from spec.md, plan.md, tasks.md and the project constitution. It builds internal semantic models (requirements inventory, user stories, task coverage map, constitution rules) and runs focused detection passes for duplication, ambiguity, underspecification, constitution alignment, coverage gaps, and inconsistencies. Findings are prioritized by severity and returned as a structured Markdown report and next-action recommendations.
When to use it
- After speckit-tasks has produced a complete tasks.md and before any implementation work
- When you need a read-only verification that tasks cover requirements and respect the constitution
- To catch ambiguous or conflicting requirements early in the workflow
- When preparing for a design or implementation handoff to engineering
- Prior to running speckit-implement to avoid wasted work
Best practices
- Run the prerequisite check script from repo root so FEATURE_DIR and docs are discovered reliably
- Limit focus areas in the request to help token-efficient analysis (e.g., security, performance)
- Address CRITICAL constitution violations before proceeding to implementation
- Use the remediation offer to get suggested edits, then apply changes manually after approval
- Re-run analysis after spec or task changes to keep IDs and counts deterministic
Example use cases
- Validate that every functional requirement in spec.md has at least one corresponding task in tasks.md
- Detect conflicting stack or framework choices between plan.md and spec.md
- Find ambiguous non-functional requirements (e.g., “fast”, “secure”) lacking measurable acceptance criteria
- Surface tasks that reference files or components absent from the spec/plan
- Produce a compact report for stakeholder review before developer handoff
FAQ
No. The skill is strictly read-only and will never write or alter repository files.
What happens if required documents are missing?
The skill aborts with a clear error instructing you to run the missing prerequisite skill or script; it will not proceed without spec.md, plan.md, and tasks.md present.