generate-verified-docs_skill

This skill generates audit-grade, hallucination-free documentation from a codebase feature with traceability and verifiable code references.
  • TypeScript

1.7k

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 quran/quran.com-frontend-next --skill generate-verified-docs

  • SKILL.md6.9 KB

Overview

This skill generates audit-grade, hallucination-free documentation for a specific codebase feature with full traceability to source lines. It produces verified technical and non-technical views, diagrams, execution flows, state machines, and a confidence report, halting if any claim cannot be tied to code. Use it to produce compliance-ready specs and developer-facing documentation that reference exact file:line evidence.

How this skill works

I parse a clear feature name and description, locate concrete entry points in the codebase, and map related artifacts and dependencies. I trace execution paths, extract explicit states and transitions, detect failure and edge cases, and build diagrams. Every factual claim is linked to file:line; if verification fails at any gate I stop and produce a failure report with partial outputs.

When to use it

  • Document a new or existing feature for audits or compliance
  • Create technical specs with provable code traceability
  • Produce non-hallucinated docs for critical systems (payments, auth, data pipelines)
  • Prepare developer onboarding materials tied to implementation
  • Validate that documentation matches the current codebase before release

Best practices

  • Provide a concise, specific feature name and short description before running the skill
  • Ensure the tool has access to the repository root and TypeScript sources
  • Narrow ambiguous scopes up-front to avoid scope rejection
  • Run on a reproducible commit or tag to keep references stable
  • Review failure reports and supply missing context or broadened scope when asked

Example use cases

  • Generate verified docs for "User Registration" with API routes and UI flows
  • Produce an execution flow and state machine for a job processing pipeline
  • Create an appendix of code references for a security audit
  • Detect and document missing error handling and edge cases in an order service
  • Produce mermaid diagrams that strictly match traced flows for architecture review

FAQ

A clear feature name and short description plus access to the codebase root. Ambiguous descriptions will prompt clarifying questions.

What happens if a claim can’t be verified?

The process halts at the enforcing gate, preserves partial outputs, and returns a failure report listing missing references and suggested recovery steps.

Can this run incrementally on partial codebases?

Yes, but any unverifiable entry points will be reported and may force a halt unless you accept partial coverage.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational