Formalize proven patterns from high-scoring research scenarios

[Copilot]

Extracted and documented 5 proven patterns from research scenarios that consistently scored 5.0/5.0, making them discoverable and reusable for future workflow development.

Changes

• Created .github/agents/patterns/proven-patterns.md (628 lines)

  • Complete catalog of 5 high-impact patterns with decision tree
  • Each pattern: when to use/not use, code examples (<20 lines), real-world examples, anti-patterns
  • Patterns: repo-memory baselines, noop for healthy states, rate limiting PRs, multi-phase pipelines, duplicate prevention

• Updated .github/agents/developer.instructions.md

  • Added "Proven Workflow Patterns" section with quick reference
  • Pattern selection guide and decision tree
  • Links to comprehensive catalog

• Created test workflow .github/workflows/test-patterns.md

  • Demonstrates combined use of 3 patterns (repo-memory, noop, duplicate prevention)
  • Validates patterns compile correctly

Pattern Examples

Monitoring with duplicate prevention:

tools:
  cache-memory: true
safe-outputs:
  create-issue:
    max: 3

# Hash issue content, check 30-day cache before creating
HASH=$(echo -n "$title$body" | sha256sum | cut -d' ' -f1)
if jq -e --arg hash "$HASH" '.issues[] | select(.hash == $hash)' "$CACHE_FILE"; then
  echo "⏭️ Skipping duplicate"
else
  create_issue && cache_hash
fi

Performance tracking with baselines:

tools:
  repo-memory:
    branch-name: memory/baselines
    create-orphan: true

# Compare against historical baseline in repo-memory
if [ "$CURRENT" -gt "$(echo "$BASELINE * 1.2" | bc)" ]; then
  echo "⚠️ Regression: ${CURRENT}ms vs ${BASELINE}ms"
fi

Impact

Future workflows can reference patterns catalog during design, reducing reinvention and improving quality scores through proven approaches.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>[plan] Formalize proven patterns from high-scoring research scenarios</issue_title>
<issue_description>## Objective

Extract and formalize proven patterns from high-scoring research scenarios into agent knowledge base, ensuring future workflows benefit from successful patterns.

Context

Research identified 5 high-impact patterns that consistently scored 5.0/5.0. These should be explicitly documented and emphasized in agent instructions.

High-Value Patterns to Formalize

1. repo-memory for baselines (Used in 4/11 scenarios)

   • Pattern: Store baseline metrics, compare new runs, detect regressions
   • Example: API performance monitoring (BE-2) - 30-day historical tracking
   • When to use: Performance tracking, visual regression, flaky test analysis

2. noop for healthy states (Prevents issue spam)

   • Pattern: Only create issues when problems detected, use noop otherwise
   • Example: Deployment monitoring (DO-1) - Only alert on failures
   • When to use: Monitoring workflows that run frequently

3. Rate limiting for automated PRs (Team-friendly automation)

   • Pattern: Limit automated PR creation to N per day with CVSS prioritization
   • Example: Security scanner (DO-2) - Max 3 PRs/day
   • When to use: Any workflow that creates PRs automatically

4. Multi-phase analysis pipelines (Complex statistical analysis)

   • Pattern: Data collection → Analysis → Historical tracking → Reporting
   • Example: Flaky test analyzer (QA-2) - 6-phase pipeline with scoring algorithm
   • When to use: Workflows requiring sophisticated analysis

5. Smart duplicate prevention (30-day cache window)

   • Pattern: Hash issues/incidents, check cache before creating
   • Example: Deployment monitoring (DO-1), Security scanner (DO-2)
   • When to use: Monitoring that detects recurring problems

Approach

1. Create patterns reference section in agent instructions
2. Document each pattern with:
   • Description and benefits
   • Code example (10-15 lines)
   • When to use / when not to use
   • Link to source scenario from research
3. Add decision tree for pattern selection
4. Update agent to proactively suggest relevant patterns based on scenario type

Files to Modify

• Update: .github/instructions/developer.instructions.md (add patterns reference section)
• Create: .github/instructions/patterns/proven-patterns.md (detailed pattern catalog)

Acceptance Criteria

• All 5 patterns documented with working code examples
• Each pattern includes "When to use" and "When NOT to use" guidance
• Decision tree helps agent select appropriate patterns
• Examples are copy-paste-ready (under 20 lines each)
• Agent instructions reference patterns section
• Tested by generating 3 workflows that leverage these patterns

Expected Impact

Ensures future workflows benefit from proven high-scoring patterns, improving average quality score and reducing reinvention of successful approaches.
Related to #10285

AI generated by Plan Command for discussion #10248

Comments on the Issue (you are @copilot in this section)

• Fixes [plan] Formalize proven patterns from high-scoring research scenarios #10288

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.