SkillVaultskills Browse all 1,000+ skills

Career · Version 1.3.0 · Reviewed 2026-08-02

Pull Request Review Learning Coach

Improve review feedback learning plan and current-diff learning card with evidence, explicit trade-offs, and a verification plan.

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

Builds short, evidence-backed learning exercises from a developer review history and the code they are changing now. It grounds the decision in review comments, resolved diffs, current changes, repository conventions, and developer-selected learning goals and explicitly prevents generic advice disconnected from the current diff or feedback quoted without the resolution that made it useful.

₹199 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 feedback learning plan
  • Current-diff learning card
  • Recurring mistake reduction

How Pull Request Review Learning Coach works

You provide

Current artifact, target level, and verifiable outcomes

It inspects

Evidence strength and scope honesty behind review feedback learning plan

It decides

A current-diff learning card rewrite with no invented achievements

You verify

Every claim survives a skeptical follow-up question

What it checks first

Pull Request Review Learning Coach builds short, evidence-backed learning exercises from a developer review history and the code they are changing now. It grounds the decision in review comments, resolved diffs, current changes, repository conventions, and developer-selected learning goals and explicitly prevents generic advice disconnected from the current diff or feedback quoted without the resolution that made it useful. Use it when the work involves Review feedback learning plan, Current-diff learning card, Recurring mistake reduction.

  1. Whether each claim describes a responsibility or a measurable outcome with the author's actual scope.
  2. Whether the evidence would survive a reference check by someone who was there.
  3. The gap between the target level's expectations and what the artifact demonstrates.
  4. Whether impact is attributed to the individual, the team, or the system, and which is honest.

Failure modes it recognizes

  • Metrics quoted without a baseline or a denominator, which a reviewer discounts entirely.
  • Team outcomes written in the first person, which fails the first probing question.
  • A packet listing activity rather than the decisions the person owned.
  • Scope inflation that is contradicted by the person's own examples during interview.

Answers it will reject

  • Inventing or rounding metrics to strengthen a claim, which is unrecoverable once challenged.
  • Optimizing for keyword matching rather than for the evidence a reviewer looks for.
  • Treating seniority as tenure rather than as demonstrated scope of judgment.

Decision rules it applies

  • Never state an outcome the person cannot substantiate under questioning.
  • Convert responsibilities into outcomes with a baseline, a change, and the person's specific role.
  • Name the largest credibility gap first; polish is worthless if the evidence is thin.

Evidence it asks for

  • For each claim, ask what a skeptical reviewer would request as proof.
  • Compare the artifact against the published expectations for the target level.
  • Rehearse the two questions most likely to expose an overstated claim.

The method inside

  1. Map the artifact, actors, boundaries, and invariants relevant to review feedback learning plan.
  2. Trace concrete failure or abuse paths for current-diff learning card; do not report checklist items without a mechanism.
  3. Prioritize recurring mistake reduction findings by impact, likelihood, confidence, and cost of correction.
  4. Recommend the smallest defensible change, then define how an independent reviewer can verify it.

Deliverables

  • Review feedback learning plan assessment
  • Current-diff learning card decision and action plan
  • Recurring mistake reduction verification checklist

Evidence requirements

  • Current resume or career artifact
  • Target role and job description
  • Verified scope, outcomes, and metrics

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 learning coach to our current review feedback learning plan work. We need a concrete decision, bounded changes, and evidence that the result is correct.

Expected output

Start with review comments, resolved diffs, current changes, repository conventions, and developer-selected learning goals. The highest-risk failure is generic advice disconnected from the current diff or feedback quoted without the resolution that made it useful. Teach one recurring mechanism at a time using an anonymized example, correction, and immediate practice check. Verify the result by tracking whether the same review category recurs in later pull requests under comparable conditions.

Boundaries and compatibility

Ideal for

  • Review feedback learning plan: produce a decision or artifact grounded in supplied evidence.
  • Current-diff learning card: produce a decision or artifact grounded in supplied evidence.
  • Recurring mistake reduction: produce a decision or artifact grounded in supplied evidence.

Out of scope

  • Inventing employers, metrics, projects, or credentials
  • Guaranteeing interviews or hiring outcomes

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.