constitution-validation_skill

This skill helps you create and validate project constitutions by discovering patterns in your codebase and generating governance rules.
  • Shell

168

GitHub Stars

3

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 rsmdt/the-startup --skill constitution-validation

  • reference.md7.8 KB
  • SKILL.md9.9 KB
  • template.md3.0 KB

Overview

This skill creates and validates project constitutions by discovering real codebase patterns and generating governance rules. It produces a project-specific CONSTITUTION.md from a template and verifies code against those rules during implementation and review. The approach emphasizes discovery, user confirmation, and a three-tier (L1/L2/L3) rule severity model.

How this skill works

First, the skill explores the repository using globbing and content reads to identify frameworks, architecture, security patterns, and conventions. It replaces [NEEDS DISCOVERY] placeholders with concrete rules derived from observed patterns, writes the constitution to the project root, and can run validation mode to apply rules and report violations. Validation parses constitution rules, globs target files, executes regex or LLM semantic checks, and produces a prioritized compliance report.

When to use it

  • When creating a new governance constitution based on the actual codebase
  • When validating code against an existing constitution during review
  • When updating constitution rules after adding frameworks or new patterns
  • During implementation to enforce blocking rules (L1/L2) and run autofixes
  • When auditing security-critical or architectural compliance before release

Best practices

  • Always run discovery against the real repository; avoid assumptions about frameworks
  • Present discovered patterns and proposed rules to users and require confirmation
  • Use L1 only for objectively fixable or security-critical issues to enable autofix
  • Reserve L2 for architecture or judgment calls that must block progress until fixed
  • Keep L3 advisory rules lightweight and optional to reduce noise in CI

Example use cases

  • Generate a CONSTITUTION.md that enforces environment variable usage instead of hardcoded secrets
  • Validate a pull request to block merges if L2 architecture rules are violated
  • Scan the codebase to detect eval() calls and auto-fix or flag them as L1
  • Audit repository boundaries to flag cross-package relative imports as L2
  • Provide a compliance report summarizing L1/L2/L3 findings for stakeholders

FAQ

The skill reports that no constitution was found and proceeds with creation if requested; validation mode will skip and notify.

How are autofixes applied?

Only L1 rules are eligible for deterministic autofix; the skill performs replacements or refactors it can do safely and reports changes for review.

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