← Back to Prompts

V2 POOL PROMPTS QUICK REFERENCE REVISED

Documentation for V2_POOL_PROMPTS_QUICK_REFERENCE_REVISED from the Dspex repository.

V2 Pool Implementation - Quick Reference Prompts (REVISED)

🚀 Quick Start Prompts

Get Current Status

Show me the current V2 pool implementation status.

First, read these files:
1. Read CLAUDE.md for progress tracking
2. Check current git status: git status
3. Check recent commits: git log --oneline -5

Then show me:
1. Current implementation phase
2. Last completed task
3. Any uncommitted changes
4. Which phase we're in

Resume Work

I need to resume V2 pool work.

First, gather context:
1. Read the last session notes from CLAUDE.md
2. Check uncommitted changes: git diff
3. Read current phase design doc section
4. Check failing tests: mix test 2>&1 | grep -A5 "test.*failure"

Show me:
1. Where we left off
2. Current work state
3. Next immediate task
4. Any blockers to resolve

🔧 Common Implementation Tasks

Implement a Function

Implement [function_name] from Design Doc [X].

First, prepare by reading:
1. Read docs/V2_POOL_TECHNICAL_DESIGN_[X]_*.md for the specification
2. Check existing patterns: grep -r "similar_function" lib/
3. Review related tests: ls test/**/[related]_test.exs

Then implement:
1. Function signature and docs
2. Core logic
3. Error handling
4. Logging/telemetry
5. Tests

Show me complete implementation.

Fix a Test

Fix the failing test at [file:line].

First, understand the failure:
1. Run: mix test [file]:[line] --trace
2. Read the test implementation
3. Read the code being tested
4. Check recent changes: git log -p -- [file]

Fix approach:
1. Identify root cause
2. Determine if test or code needs fixing
3. Implement fix
4. Verify fix works

Show me the fix and verification.

Add Error Handling

Add error handling to [module/function].

First, analyze error cases:
1. Read the function implementation
2. Check docs/V2_POOL_TECHNICAL_DESIGN_4_ERROR_HANDLING.md
3. Find similar error handling: grep -r "PoolErrorHandler" lib/

Implement:
1. Identify all error cases
2. Use PoolErrorHandler.wrap_pool_error
3. Add appropriate logging
4. Include telemetry events
5. Test error scenarios

Show me updated code with error handling.

🧪 Testing Commands

Run Tests by Layer

# First check which layer based on current phase or feature

# Layer 1 (Mock) - Unit tests
mix test --only layer_1

# Layer 2 (Bridge Mock) - Integration with mock bridge
TEST_MODE=bridge_mock mix test --only layer_2

# Layer 3 (Full Integration) - Real Python processes
TEST_MODE=full_integration mix test --only layer_3

# Specific module
mix test test/path/to/module_test.exs

# With coverage
mix coveralls.html

Run Specific Test Types

# First determine test type needed

# Unit tests only
mix test test/unit/

# Integration tests
mix test test/integration/

# Performance tests
mix test --only performance

# Chaos tests
mix test --only chaos

# Single test
mix test test/file.exs:LINE

🐛 Debugging Prompts

Debug Pool Issues

Debug pool issue: [describe problem].

First, gather information:
1. Check pool processes: :observer.start()
2. Read recent logs: grep -n "ERROR\|WARN" log/
3. Check worker states in observer
4. Review telemetry: check ETS tables

Debug approach:
1. Add IO.inspect at key points
2. Use :dbg for function tracing
3. Check telemetry events
4. Create minimal reproduction

Show me findings and fix approach.

Trace Execution

Trace execution of [operation].

First, identify trace points:
1. Find the entry function
2. Map the call flow
3. Identify state changes

Add tracing:
1. IO.inspect(binding(), label: "POINT_NAME")
2. Use :dbg.tracer()
3. Add telemetry events
4. Create focused test

Show me trace results and analysis.

📊 Performance Prompts

Check Performance

Check performance of [component].

First, measure current state:
1. Find or create benchmark: ls bench/
2. Check existing metrics: read telemetry data
3. Review design targets

Measure:
1. Run benchmarks with Benchee
2. Check current metrics
3. Compare with baseline
4. Profile with :fprof if needed

Show me performance analysis.

Optimize Component

Optimize [component] performance.

First, profile the component:
1. Measure current performance
2. Run :fprof or :eprof
3. Check algorithmic complexity
4. Review data structures

Optimize:
1. Identify bottleneck
2. Implement optimization
3. Measure improvement
4. Ensure no regression

Show me before/after metrics.

🔍 Code Search Prompts

Find Pattern

Find all instances of [pattern].

Search commands:
# Text search
grep -r "pattern" lib/ test/

# Function definitions
grep -r "def pattern" lib/

# Module references
grep -r "alias.*Pattern\|Pattern\." lib/

# Test examples
grep -r "pattern" test/

Show me all occurrences with context.

Find code related to [feature].

First, identify search terms:
1. Module names
2. Function names
3. Key variables
4. Error messages

Search:
1. grep -r "feature\|Feature" lib/
2. Check similar modules
3. Look in tests
4. Review docs

Show me related code organized by purpose.

📝 Documentation Prompts

Update Documentation

Update documentation for [component].

First, check what needs updating:
1. Read current @moduledoc in file
2. Check function @doc strings
3. Review CLAUDE.md mentions
4. Check README if applicable

Update:
1. Module documentation
2. Function documentation
3. Add examples
4. Update CLAUDE.md
5. Update guides if needed

Show me documentation updates.

Create Runbook

Create runbook for [scenario].

First, understand the scenario:
1. When does it occur?
2. What are symptoms?
3. What's the impact?

Document:
1. Scenario description
2. Detection methods
3. Step-by-step resolution
4. Verification steps
5. Prevention measures

Show me complete runbook.

🚢 Deployment Prompts

Prepare Release

Prepare for release of current work.

First, check readiness:
1. Run full test suite: mix test
2. Check coverage: mix coveralls
3. Run formatter: mix format
4. Check credo: mix credo

Prepare:
1. Update version if needed
2. Create release notes
3. Document breaking changes
4. Update migration guide

Show me release checklist status.

Validate Deployment

Validate deployment readiness.

Check:
□ All tests passing?
□ Performance targets met?
□ Documentation complete?
□ Rollback plan ready?
□ Monitoring configured?
□ Feature flags set?

Show me validation results.

💡 Architecture Decisions

Evaluate Options

Evaluate options for [decision].

First, understand context:
1. Read relevant design doc section
2. Check existing patterns
3. Review constraints

Evaluate:
1. List all options
2. Compare pros/cons
3. Consider our design principles
4. Check performance impact
5. Assess maintenance burden

Show me recommendation with rationale.

Review Design

Review design of [component].

First, read implementation:
1. Read the module/function
2. Check against design doc
3. Review test coverage

Assess:
□ Follows our patterns?
□ Error handling complete?
□ Edge cases covered?
□ Testable design?
□ Performance acceptable?

Show me review findings.

🛟 Emergency Prompts

System Down

Pool system is down!

First, diagnose:
1. Check supervisor: Process.whereis(DSPex.PythonBridge.Supervisor)
2. Look for crashes: ls log/crash.log*
3. Check system resources: File.read!("/proc/meminfo")
4. Review recent changes: git log --oneline -10

Recover:
1. Restart supervisor if needed
2. Check pool configuration
3. Verify Python processes
4. Use fallback adapter

Show me diagnosis and recovery steps.

Performance Crisis

Performance is severely degraded!

First, measure:
1. Check current metrics vs normal
2. Look at pool utilization
3. Check error rates
4. Monitor resource usage

Mitigate:
1. Identify bottleneck
2. Scale pool if needed
3. Enable circuit breakers
4. Reduce timeout values
5. Switch to degraded mode

Show me immediate actions to take.

📋 Checklists

Before Committing

Pre-commit checklist:

First run these checks:
□ mix test - All tests pass?
□ mix format --check-formatted - Code formatted?
□ mix credo - No issues?
□ grep -r "IO.inspect\|TODO" lib/ - No debug code?
□ git diff - Changes look correct?

Show me any issues found.

End of Day

End of day checklist:

Tasks to complete:
□ Commit current work
□ Update CLAUDE.md with progress
□ Push to branch: git push
□ Note tomorrow's first task
□ Clean up any temp files

Show me what's left to do.

🎯 Phase-Specific

Current Phase Check

What phase am I in?

First, check indicators:
1. Read CLAUDE.md header section
2. Check recent commits: git log --grep="Phase"
3. Look at current branch name
4. Check which design doc we're using

Show me:
1. Current phase
2. Phase progress
3. Next milestone

Phase Transition

Ready to move to next phase?

Validation checklist:
□ All current phase tests pass?
□ Documentation complete?
□ No critical issues?
□ Performance acceptable?
□ Design doc tasks done?

Show me validation results and next phase prep.

Remember:

  • Always read relevant files first before making changes
  • Check existing patterns before implementing new code
  • Run tests after changes
  • Update documentation as you go
  • For detailed implementation, refer to the phase-specific prompt files