convention-mining

Convention Mining

Surface implicit coding conventions from merged PR review comments, then verify and harden the results. This skill covers two phases: mining conventions from review data, and auditing convention skills for accuracy.

Workflow

Phase	When to Use	Guide
Mining	Extracting conventions from PR review history for the first time	mining.md
Auditing	Verifying a mined skill is factually correct before merging, or hardening an existing conventions skill	auditing.md

Mining produces a conventions document or skill. Auditing verifies that the output is accurate — quotes are real, code examples match the actual codebase, and nothing contradicts existing docs. Run auditing after mining, or independently on any conventions skill that needs review.

Phase Routing

Route to mining.md when the user wants to:

• Extract conventions from PR review history
• Understand what reviewers enforce on a repo
• Create a conventions guide, skill, or CLAUDE.md additions
• Onboard to a new codebase's review culture

Route to auditing.md when the user wants to:

• Audit or review a mined conventions skill
• Verify that quoted review comments are real
• Check code examples against the actual codebase
• Find gaps, overlaps, or contradictions with existing docs
• Harden a conventions skill before merging

References

Detailed command references and templates are in the references/ directory:

• references/gh-commands.md — GitHub CLI commands for fetching PR review data
• references/output-templates.md — Templates for conventions documents, skills, and CLAUDE.md additions
• references/verification-techniques.md — Commands and procedures for verifying quotes, finding gaps, and checking code examples