---

name: implementing-performance-fixes description: "[PHASE 4] Implements and validates fixes for verified theories. Entry: PERFORMANCE_THEORIES.md has a verified theory. Outputs: Code fix, BENCHMARK_BASELINE.md new column, theory marked fixed/fix-failed. Next: Loop to PHASE 2 if performance gaps remain, otherwise done."

Implementing Performance Fixes

Transforms verified performance theories into working fixes with measured validation.

Core Principle

Verified theories identify the problem. This skill implements and validates the solution.

Prerequisites

• REQUIRED: PERFORMANCE_THEORIES.md with at least one theory status verified
• REQUIRED: BENCHMARK_BASELINE.md with benchmark descriptions and execution commands
• REQUIRED: Theory verification evidence from verifying-performance-theories

Quick Start

Step 1: Load the most critical verified theory from PERFORMANCE_THEORIES.md

Step 2: Design fix approach (see WORKFLOW.md Phase 2)

Step 3: Implement the fix

Step 4: Run 3 relevant benchmarks to validate (select based on benchmark focus in BENCHMARK_BASELINE.md)

Step 5: Based on results:

• Improved → Keep fix, run all benchmarks, update files
• Degraded → Revert fix, update theory status

Workflow Overview

Copy1. Load verified theory with evidence
2. Design fix (architectural changes allowed if cost-benefit fits)
3. Implement fix
4. Run 3 relevant benchmarks
5a. If improved → run all benchmarks → update BENCHMARK_BASELINE → mark theory "fixed"
5b. If degraded → revert → mark theory "fix-failed"

Benchmark Selection

Select 3 benchmarks from BENCHMARK_BASELINE.md based on their documented focus:

1. Read each benchmark's Rationale field
2. Match rationale to the code path affected by your fix
3. Prioritize benchmarks that exercise the fixed code

File Updates

BENCHMARK_BASELINE.md

Add new column to results table:

Copy| Benchmark | Baseline | After Fix #N |
|-----------|----------|--------------|
| queens    | 32ms     | 24ms         |

PERFORMANCE_THEORIES.md

Update theory status:

Copy**Status**: fixed  <!-- was: verified -->
**Fix Applied**: [Date] - [Brief description of fix]
**Performance Impact**: [X% improvement on benchmark Y]

Or if fix failed:

Copy**Status**: fix-failed  <!-- was: verified -->
**Fix Attempted**: [Date] - [What was tried]
**Reason**: [Why it didn't improve performance]

Status Values

Integration

Predecessor: verifying-performance-theories → Provides verified theories

Successor: Loop back to generating-performance-theories if more issues exist

Workflow Position

Copy1. [establishing-benchmark-baseline] → Create baseline (PHASE 1)
2. [generating-performance-theories] → Generate theories (PHASE 2)
3. [verifying-performance-theories]  → Verify with tools (PHASE 3)
4. [implementing-performance-fixes]  → THIS SKILL (PHASE 4)
5. Loop to step 2 if performance gaps remain

See WORKFLOW.md for detailed phase instructions.