- Home
- Skills
- Vladm3105
- Aidoc Flow Framework
- Doc Spec Reviewer
doc-spec-reviewer_skill
- Python
9
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 vladm3105/aidoc-flow-framework --skill doc-spec-reviewer- SKILL.md18.2 KB
Overview
This skill provides comprehensive content review and quality assurance for Technical Specification (SPEC) documents. It validates YAML structure, checks REQ coverage and interface definitions, detects upstream drift, and produces a prioritized issue list plus a review score. The tool enforces mandatory drift cache behavior and flags blocking structural violations before deeper checks.
How this skill works
The reviewer loads a SPEC YAML, verifies nested folder and 13-section completeness, and runs nine focused checks (structure, REQ coverage, interfaces, thresholds, data models, error handling, placeholders, naming, drift). It computes SHA-256 hashes for files and sections to detect upstream drift, updates a mandatory .drift_cache.json atomically, and outputs a review score with categorized issues and auto-fix where safe.
When to use it
- Immediately after automated SPEC generation (post doc-spec-autopilot)
- After manual edits to a SPEC before downstream tasks
- As a pre-check before generating TASKS or TSPEC artifacts
- Regularly for periodic quality assurance of existing SPECs
- When traceability to REQ/CTR must be validated and drift tracked
Best practices
- Always run after any change to REQ or CTR documents to capture drift early
- Keep the .drift_cache.json under version control and ensure cache_version compatibility
- Resolve blocking structure errors (nested folder, filename) before other reviews
- Treat placeholder detections ([TODO]/[TBD]) as highest-priority manual work items
- Review and accept cached updates only after human verification of substantive upstream changes
Example use cases
- Final QA gate before TASKS/TSPEC generation to ensure implementation readiness
- Catching missing REQ implementations or unmapped acceptance criteria in a SPEC
- Detecting changed upstream API contracts and flagging section-level drift for review
- Validating that interface and method signatures are complete and implementation-ready
- Ensuring threshold registry values match upstream requirements and SLAs
FAQ
A missing or invalid drift cache triggers an error (REV-D006) and initializes a new cache; the review still runs but you must resolve the cache state and commit it to enable accurate incremental drift detection.
Are any fixes applied automatically?
The skill auto-fixes only safe, structural or formatting issues where deterministic changes are available. Content issues, placeholders, and upstream drift are flagged for manual review.