
performance-forensics
Measure and diagnose performance with reproducible evidence.
- version
- 0.1.4
- GitHub stars
- 347
Install
/plugin install performance-forensics@khanelinix-ai-toolsRun it inside Claude Code after adding the marketplace with /plugin marketplace add khaneliman/khanelinix.
Contents
Discovered from sourceSkills 1
Performance Forensics
Use this skill for empirical performance diagnosis and optimization. Performance engineering requires measurement artifacts instead of speculative code inspection.
For memory leaks, out-of-memory errors, and heap fragmentation, use
memory-profiler.
Modes of Operation
Select the mode that matches the task trigger:
- Live process diagnosis: Inspect active spinning processes, thread contention, or I/O stalls. Read references/runtime-forensics.md.
- Static trace diagnosis: Analyze captured profiles, CPU dumps, spindumps, and traces. Read references/trace-forensics.md.
- One-off optimization: Resolve a specific latency, throughput, or CPU bottleneck. Read references/perf-issue.md.
- Iterative hillclimbing: Optimize a metric across multiple experimental cycles. Read references/hillclimb.md.
Diagnosis Workflow (Read-Only Default)
Diagnosis tasks are read-only by default. Do not mutate production code during diagnosis unless requested.
- Capture evidence. Obtain a CPU profile, trace, or execution log using a realistic workload.
- Reduce artifacts. Query the profile to identify the dominant bottleneck or hot call tree.
- Confirm the mechanism. Verify the bottleneck on the running target before proposing changes.
- Attribute to source. Map the hotspot to the exact file, symbol, and line number.
- Report findings. Provide the diagnosis, source location, and artifact paths.
Optimization Workflow
When authorized to mutate code, follow this measurement loop:
- Ground the workload. Define a representative workload matching real input distributions.
- Establish baseline. Capture baseline metrics and profile artifacts. Save the artifact path.
- Formulate single hypothesis. Pick a targeted strategy from the eight strategy families in references/perf-issue.md.
- Apply isolated change. Modify only the code related to the selected hypothesis.
- Measure post-fix result. Run the identical workload. Compare the new artifact against baseline.
- Evaluate outcome. Keep confirmed improvements. Restore only experiment changes owned by this run after regressions or neutral results. Flag inconclusive measurements.
- Iterate or stop. For sustained optimization, follow the stop rules in references/hillclimb.md.
Reporting Contract
For diagnosis, return the workload, baseline artifact, source attribution, and status. For authorized optimization, also return:
- Workload description and parameters.
- Post-fix measurement number and artifact path.
- Net delta percentage or throughput difference.
- Status classification: improvement, regression, or inconclusive.
Sponsors
Details
- Marketplace
- khaneliman/khanelinix
- Source
- modules/common/ai-tools/skills/performance-forensics
- Author
- khaneliman
- Category
- Developer Tools
More from khaneliman/khanelinix
All 43- planning-with-filesv0.1.9
- program-orchestrationv0.1.5
- recallv0.1.7
- requirements-interviewv0.1.4
- researchv0.1.3
- rust-toolkitv0.1.6
- sarif-toolkitv0.1.5
- security-toolkitv0.1.7
- show-me-your-workv0.1.7
- skill-creatorv0.1.8
- software-engineeringv0.1.9
- tddv0.1.4
- technical-writingv0.2.0
- to-specv0.1.1
- to-ticketsv0.1.1
- typescript-best-practicesv0.1.6
- unslopv0.1.7
- verification-harnessv0.1.5
- whyv0.1.6
- ai-tools-architectv0.1.21
- architectv0.1.7
- arenav0.1.5
- bevy-toolkitv0.1.6
- blast-radiusv0.1.6