SkillVaultskills Browse all 1,000+ skills

Research · Version 1.0.0 · Reviewed 2026-08-02

Pull Request Review Knowledge Miner

Produce defensible evidence for review history mining and institutional rule synthesis with evidence, explicit trade-offs, and a verification plan.

4 method steps 4 documented failure modes 4 diagnostic checks 7 quality gates

Extracts recurring engineering rules and rationale from resolved pull request discussions without turning preferences into policy. It grounds the decision in merged pull requests, complete review threads, final diffs, reviewer roles, recurrence, and existing guidance and explicitly prevents mining comments without their resolution and codifying advice that maintainers later rejected or narrowed.

₹299 one-time

Get this skill archive

Install in your AI coding tool

SkillVault packages this skill in the open Agent Skills format for five leading coding tools.

What this skill helps you do

  • Review history mining
  • Institutional rule synthesis
  • Reviewer authority analysis

How Pull Request Review Knowledge Miner works

You provide

Sources, method notes, and the question being answered

It inspects

Evidence strength and negative cases for review history mining

It decides

A institutional rule synthesis synthesis with limitations named

You verify

Findings reproducible from the cited evidence alone

What it checks first

Pull Request Review Knowledge Miner extracts recurring engineering rules and rationale from resolved pull request discussions without turning preferences into policy. It grounds the decision in merged pull requests, complete review threads, final diffs, reviewer roles, recurrence, and existing guidance and explicitly prevents mining comments without their resolution and codifying advice that maintainers later rejected or narrowed. Use it when the work involves Review history mining, Institutional rule synthesis, Reviewer authority analysis.

  1. Sample size, selection method, and who was excluded, since these bound every conclusion.
  2. Whether a finding is supported by observed behavior or by what participants said they would do.
  3. Disconfirming cases, which are the fastest test of whether a theme is real.
  4. Whether the question the research answers is the question the decision actually needs.

Failure modes it recognizes

  • Themes derived from the loudest participants rather than from the sampled distribution.
  • Leading questions producing agreement that disappears under neutral phrasing.
  • Survivorship bias from studying only current customers or successful accounts.
  • Coding drift where the same excerpt would be categorized differently later in the process.

Answers it will reject

  • Reporting percentages from a small qualitative sample, which implies precision that does not exist.
  • Treating a synthesized theme as a finding without the excerpts that support it.
  • Stopping at the first pattern that confirms the pre-existing hypothesis.

Decision rules it applies

  • State the sample, the method, and the limitation alongside every finding.
  • Actively seek disconfirming evidence before accepting a theme.
  • Separate what was observed, what was reported, and what was inferred.

Evidence it asks for

  • Attach verbatim excerpts to each theme so a reader can audit the interpretation.
  • Check coding consistency by re-coding a subset and comparing.
  • Report counts with denominators rather than percentages alone.

The method inside

  1. Define the research question and unit of analysis
  2. Create a transparent coding or extraction framework
  3. Preserve source traceability and negative evidence
  4. Separate findings, interpretation, limitations, and applicability

Deliverables

  • Review history mining evidence table
  • Institutional rule synthesis findings with negative cases
  • Reviewer authority analysis limitations and next-research plan

Evidence requirements

  • Source documents, transcripts, data, and research question
  • Sampling method, population, and collection context
  • Known limitations, contradictory cases, and analysis criteria

Quality gates

  • Every material claim traces to supplied evidence or is labeled as a hypothesis.
  • The response follows the declared deliverable contract.
  • No execution, access, measurement, or verification is invented.
  • Secrets and personal data are redacted rather than repeated.
  • The user receives a concrete independent verification step.
  • The relevant failure modes in this domain were considered rather than only the reported symptom.
  • No listed anti-pattern was recommended as a solution.

Example task

Input

Apply the pull request review knowledge miner to our current review history mining work. We need a concrete decision, bounded changes, and evidence that the result is correct.

Expected output

Start with merged pull requests, complete review threads, final diffs, reviewer roles, recurrence, and existing guidance. The highest-risk failure is mining comments without their resolution and codifying advice that maintainers later rejected or narrowed. Require repeated evidence or authoritative confirmation and preserve the scope and counterexamples of every synthesized rule. Verify the result by backtesting proposed knowledge against the source pull requests and obtaining owner approval before publication.

Boundaries and compatibility

Ideal for

  • Review history mining: produce a decision or artifact grounded in supplied evidence.
  • Institutional rule synthesis: produce a decision or artifact grounded in supplied evidence.
  • Reviewer authority analysis: produce a decision or artifact grounded in supplied evidence.

Out of scope

  • Fabricating sources, participants, or findings
  • Claiming representativeness without a sampling basis

Agent compatibility

  • GitHub Copilot Agent Skills
  • Cursor Agent Skills
  • Claude Code Skills
  • OpenAI Codex Skills
  • JetBrains Junie Skills

Tool policy: Advisory by default. No tools are assumed. If the host provides tools, use read-only evidence gathering unless the user explicitly approves a scoped write or execution action.