Skip to content
Agent PR Reviewer Skill logo

Agent PR Reviewer Skill

Evaluates PRs containing agent changes using specialized agents to analyze quality, compliance, and improvement value. Provides turbobeest with a structured recommendation to streamline human review.

SKILL.md

Full skill instructions

Agent PR Reviewer Skill

Purpose

When contributors submit PRs with agent improvements to turbobeest/​dev-system, this skill:

  1. Extracts the agent changes from the PR
  2. Validates against -01-agent-formatting/ standards
  3. Analyzes using specialized agents
  4. Summarizes changes and impact
  5. Recommends approve, request changes, or decline

Activation

/​agent-pr-reviewer pr=123

Or with full URL:

/​agent-pr-reviewer pr=https://github.com/​turbobeest/​dev-system/​pull/​123

Workflow

┌─────────────────────────────────────────────────────────────────────────┐
│                     PR REVIEW WORKFLOW                                  │
├─────────────────────────────────────────────────────────────────────────┤
│                                                                         │
│  Step 1: FETCH PR                                                       │
│  ├─ gh pr view {number} --json                                         │
│  ├─ Extract changed files                                              │
│  └─ Identify agent files (agents/​**/​*.md)                              │
│          ▼                                                              │
│  Step 2: VALIDATE STRUCTURE                                             │
│  ├─ Run agent-creation-validator on each changed agent                 │
│  ├─ Check tier compliance                                              │
│  └─ Verify required sections present                                   │
│          ▼                                                              │
│  Step 3: ANALYZE CHANGES                                                │
│  ├─ Diff analysis: what changed?                                       │
│  ├─ Knowledge sources: added/​removed/​modified?                         │
│  ├─ Instructions: improved or degraded?                                │
│  └─ Potential regressions flagged                                      │
│          ▼                                                              │
│  Step 4: AGENT EVALUATION                                               │
│  ├─ first-principles-engineer: Is this a good change?                  │
│  ├─ Domain expert (if applicable): Domain accuracy?                    │
│  └─ prd-auditor: Quality assessment                                    │
│          ▼                                                              │
│  Step 5: GENERATE RECOMMENDATION                                        │
│  ├─ Synthesize agent opinions                                          │
│  ├─ Flag concerns or conflicts                                         │
│  └─ Provide structured recommendation                                  │
│                                                                         │
└─────────────────────────────────────────────────────────────────────────┘

Output Format

╔═══════════════════════════════════════════════════════════════════════╗
║              PR REVIEW: #{number}                                      ║
║              {pr-title}                                                ║
╠═══════════════════════════════════════════════════════════════════════╣
║                                                                        ║
║  Author: {author}                                                      ║
║  Branch: {head} → {base}                                              ║
║  Files Changed: {count} ({agent-count} agent files)                   ║
║                                                                        ║
╠═══════════════════════════════════════════════════════════════════════╣
║  AGENT CHANGES                                                         ║
╠═══════════════════════════════════════════════════════════════════════╣
║                                                                        ║
║  ┌────────────────────────────────────────────────────────────────┐   ║
║  │ MODIFIED: typescript-pro.md                                     │   ║
║  ├────────────────────────────────────────────────────────────────┤   ║
║  │ + Added 3 knowledge sources                                     │   ║
║  │ ~ Modified interpretive lens                                    │   ║
║  │ ~ Updated 4 instructions                                        │   ║
║  │ - Removed 1 deprecated reference                                │   ║
║  │                                                                 │   ║
║  │ Structure: ✓ Valid                                              │   ║
║  │ Tier: Expert (unchanged)                                        │   ║
║  │ Token Δ: +120 tokens (within budget)                           │   ║
║  └────────────────────────────────────────────────────────────────┘   ║
║                                                                        ║
║  ┌────────────────────────────────────────────────────────────────┐   ║
║  │ NEW: deno-expert.md                                             │   ║
║  ├────────────────────────────────────────────────────────────────┤   ║
║  │ Category: backend-ecosystems/​javascript-runtimes               │   ║
║  │ Tier: Expert                                                    │   ║
║  │ Model: sonnet                                                   │   ║
║  │                                                                 │   ║
║  │ Structure: ✓ Valid                                              │   ║
║  │ Curation Record: ✓ Present                                      │   ║
║  │ Overlap Check: Low overlap with nodejs-expert                   │   ║
║  └────────────────────────────────────────────────────────────────┘   ║
║                                                                        ║
╠═══════════════════════════════════════════════════════════════════════╣
║  AGENT ANALYSIS                                                        ║
╠═══════════════════════════════════════════════════════════════════════╣
║                                                                        ║
║  first-principles-engineer:                                            ║
║    "The typescript-pro changes add authoritative sources from         ║
║     TypeScript's official handbook. The instruction modifications     ║
║     improve specificity without losing generality. Recommend merge."  ║
║    Confidence: HIGH                                                    ║
║                                                                        ║
║  typescript-pro (self-review of diff):                                ║
║    "New knowledge sources are accurate and high-authority.            ║
║     Instruction changes align with TypeScript 5.x patterns.           ║
║     No concerns."                                                      ║
║    Confidence: HIGH                                                    ║
║                                                                        ║
║  prd-auditor:                                                          ║
║    "Deno-expert is well-structured but missing explicit               ║
║     differentiation from nodejs-expert in Identity section.           ║
║     Minor improvement suggested."                                      ║
║    Confidence: MEDIUM                                                  ║
║                                                                        ║
╠═══════════════════════════════════════════════════════════════════════╣
║  CONCERNS                                                              ║
╠═══════════════════════════════════════════════════════════════════════╣
║                                                                        ║
║  [LOW] deno-expert could clarify differentiation from nodejs-expert   ║
║                                                                        ║
╠═══════════════════════════════════════════════════════════════════════╣
║  RECOMMENDATION                                                        ║
╠═══════════════════════════════════════════════════════════════════════╣
║                                                                        ║
║  ┌────────────────────────────────────────────────────────────────┐   ║
║  │  ✅ RECOMMEND APPROVE                                           │   ║
║  │                                                                 │   ║
║  │  • typescript-pro improvements are high-quality                │   ║
║  │  • deno-expert adds value with minor suggested tweak           │   ║
║  │  • All structure validations pass                              │   ║
║  │  • Curation records present                                    │   ║
║  └────────────────────────────────────────────────────────────────┘   ║
║                                                                        ║
║  Optional comment to author:                                           ║
║  "Consider adding explicit differentiation from nodejs-expert in      ║
║   deno-expert's Identity section (e.g., 'Deno-first approach with    ║
║   native TypeScript, secure-by-default permissions model')."          ║
║                                                                        ║
╚═══════════════════════════════════════════════════════════════════════╝

Recommendation Categories

RecommendationMeaningAction
APPROVEHigh-quality changes, ready to mergeturbobeest merges
APPROVE WITH SUGGESTIONSGood changes, minor improvements optionalMerge now or wait for tweaks
REQUEST CHANGESIssues that should be addressedAuthor revises, re-review
DECLINEDoesn't meet standards or conflicts with philosophyClose with explanation

Validation Checks

Structure Validation

  • Frontmatter complete (name, tier, model, tools, modes)
  • Identity section with interpretive lens
  • Instructions section with Always/​Mode-specific/​Never
  • Specializations defined
  • Knowledge sources documented
  • Output format specified

Quality Validation

  • Tier classification appropriate
  • Model selection justified
  • Knowledge sources authoritative
  • Instructions non-conflicting
  • Vocabulary calibrated (15-20 terms)

Contribution Validation

  • Curation record present (for new agents)
  • No duplicate agents without differentiation
  • Follows -01-agent-formatting/ standards
  • PR description explains rationale

Agent Panel

The review uses a panel of agents:

AgentRoleFocus
first-principles-engineerQuality arbiterIs this fundamentally a good change?
{domain-expert}Domain accuracyAre domain-specific details correct?
prd-auditorStandards complianceDoes it meet quality standards?
architect-reviewerStructural coherenceDoes the agent fit the ecosystem?

Integration

GitHub CLI Commands Used

# Fetch PR details
gh pr view {number} --json title,author,body,files,headRefName,baseRefName

# Get diff
gh pr diff {number}

# Add review comment (after human approval)
gh pr review {number} --approve --body "..."
gh pr review {number} --request-changes --body "..."

Signals

SignalMeaning
PR_REVIEW_STARTEDAnalysis begun
PR_VALIDATION_COMPLETEStructure checks done
PR_ANALYSIS_COMPLETEAgent analysis done
PR_RECOMMENDATION_READYReady for human review

Human Override

The skill provides a recommendation, but turbobeest makes the final decision.

After reviewing the analysis:

  • [A] Accept recommendation and merge
  • [R] Request changes (with specific feedback)
  • [D] Decline PR
  • [O] Override recommendation with rationale

All decisions are logged in .claude/​pr-reviews/ for accountability.

More Finance & Trading skills

ponytail-debt logo
Finance & Trading

ponytail-debt

Harvest every `ponytail:` comment in the codebase into a debt ledger, so the deliberate shortcuts and deferrals ponytail leaves behind get tracked instead of rotting into "later means never". One-shot report, changes nothing.

155.5K 27.9K
View
stripe-integration logo
Finance & Trading

stripe-integration

Stripe payment processing with checkout sessions, subscriptions, webhooks, and customer management.

40.2K 12.9K
View

Build comprehensive 3-5 year financial models with revenue projections, cost structures, cash flow analysis, and scenario planning for early-stage startups. Use this skill when creating financial projections, calculating burn rate or runway, modeling fundraising scenarios, or preparing investor-r...

40.2K 12K
View
stock-analysis logo
Finance & Trading

stock-analysis

Stock and cryptocurrency analysis with portfolio management and multi-dimensional scoring.

46 11.9K
View
coinglass logo
Finance & Trading

coinglass

Real-time crypto derivatives positioning, whale tracking, liquidations, and institutional ETF flows across 37 specialized tools.

27 11.5K
View

Master dbt (data build tool) for analytics engineering with model organization, testing, documentation, and incremental strategies. Use when building data transformations, creating data models, or implementing analytics engineering best practices.

40.2K 10.5K
View

Finance AI tools

C

ChainClarity explains 700+ crypto whitepapers in plain English, with layered summaries, comparisons, research tools, alerts, and a $4.99 Pro plan.

Freemium
View
Finalle.ai logo
Finance

Finalle.ai

Stay Ahead with Finalle.ai: The Ultimate Financial News and Tools Platform

Free
View

AI-powered Pine Script code generator for TradingView

Paid
View
TradeUI logo
Finance

TradeUI

Empower your investment strategies with AI signals, option flow data, and advanced chart patterns.

Paid
View
D
Finance

Danelfin

Danelfin review covering its AI stock and ETF scores, rankings, portfolio tools, current pricing, strengths, risks, and alternatives for investors.

Freemium
View
E
Finance

Ellisense

A software for comprehensive market sentiment analysis.

Paid
View

'Stock Correlations' App by Freshly.ai: Reveal Stock Patterns and Trends

Free
View