perf-theory-tester

$npx mdskill add ComposioHQ/awesome-claude-plugins/perf-theory-tester

Validate performance hypotheses through controlled single-change experiments.

  • Enables agents to test specific performance improvements with clean baselines.
  • Depends on benchmark commands and file change tracking for evidence.
  • Decides outcomes by comparing metrics against baseline after one change.
  • Delivers structured verdicts with commands, files, and delta data.

SKILL.md

.github/skills/perf-theory-testerView on GitHub ↗
---
name: perf-theory-tester
description: "Use when running controlled perf experiments to validate hypotheses."
version: 1.0.0
---

# perf-theory-tester

Test hypotheses using controlled experiments.

Follow `docs/perf-requirements.md` as the canonical contract.

## Required Steps

1. Confirm baseline is clean.
2. Apply a single change tied to the hypothesis.
3. Run 2+ validation passes.
4. Revert to baseline before the next experiment.

## Output Format

```
hypothesis: <id>
change: <summary>
delta: <metrics>
verdict: accept|reject|inconclusive
evidence:
  - command: <benchmark command>
  - files: <changed files>
```

## Constraints

- One change per experiment.
- No parallel benchmarks.
- Record evidence for each run.

More from ComposioHQ/awesome-claude-plugins

SkillDescription
add-subSubscribe to a skill event. Adds a subscription that injects context before or after a skill runs. Supports optional conditions at both insert-level (inherited by all subscriptions) and subscription-level (AND-stacked). Conditions include fileExists, gitBranch, envSet, envEquals, fileContains (with optional regex).
helpShow skill-bus overview, status, and available commands. Quick reference for all skill-bus features.
list-subsList all active skill-bus subscriptions across global and project scopes, showing merge status, insert-level and subscription-level conditions, effective condition stacking, and what would fire for each skill.
pause-subsTemporarily disable the skill bus. Quick toggle to stop all subscriptions from firing without removing them.
perf-analyzer"Use when synthesizing perf findings into evidence-backed recommendations and decisions."
perf-baseline-manager"Use when managing perf baselines, consolidating results, or comparing versions. Ensures one baseline JSON per version."
perf-benchmarker"Use when running performance benchmarks, establishing baselines, or validating regressions with sequential runs. Enforces 60s minimum runs (30s only for binary search) and no parallel benchmarks."
perf-code-paths"Use when mapping code paths, entrypoints, and likely hot files before profiling."
perf-investigation-logger"Use when appending structured perf investigation notes and evidence."
perf-profiler"Use when profiling CPU/memory hot paths, generating flame graphs, or capturing JFR/perf evidence."