perf-analyzer — Claude Code subagent: what it does, how to install
perf-analyzer — a Claude Code subagent that turns profiling results and benchmark experiments into evidence-backed recommendations structured by a strict template.
MIT ★ 2,003 updated 2026-07-26
/plugin marketplace add composio-community/awesome-claude-plugins
/plugin install perf@awesome-claude-pluginsWhat perf-analyzer does
Running on the opus model with Read and Write tools, this subagent digests raw benchmarking data. It reviews baseline metrics, profiler traces, tested hypotheses, and breaking point limits. The agent strictly treats docs/perf-requirements.md as its primary contract and must execute the perf-analyzer skill to deliver its findings.
The generated report follows a rigid structure containing summary, recommendations, abandoned, and next_steps. It highlights failed hypotheses alongside valid suggestions. Grounding is mandatory here: the subagent only cites data present in logs or code, requesting a complete test re-run whenever numbers are insufficient.
What it helps with
- Synthesize benchmark results and profiling data into structured reports
- Document failed optimization hypotheses under the abandoned section
- Verify performance findings against docs/perf-requirements.md
- Determine next testing steps based on verified log data
Example prompts for the agent
- Analyze the profiler traces and generate a report using the perf-analyzer subagent
- Review the benchmark metrics against docs/perf-requirements.md
- Synthesize the load test results and identify failed hypotheses
How to install perf-analyzer
1Plugin marketplace
/plugin marketplace add composio-community/awesome-claude-plugins/plugin install perf@awesome-claude-plugins2Manually: put the files in place
.claude/agents/perf-analyzer.mdFor all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention
Manually: put the files in place
.codex/agents/perf-analyzer.tomlCodex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/
Manually: put the files in place
.cursor/agents/perf-analyzer.mdCursor also reads .claude/agents/
Step by step, with commands for each platform: Claude Code subagents: how to create one, call it and keep context clean
Questions about perf-analyzer
How do I install perf-analyzer in Claude Code?
Add the marketplace by running /plugin marketplace add composio-community/awesome-claude-plugins. Next, install the package using /plugin install perf@awesome-claude-plugins. The agent definition is placed at .claude/agents/perf-analyzer.md.
What does the perf-analyzer subagent do?
It converts performance tests, baseline data, and profiler evidence into a structured YAML-like summary. The agent records actionable recommendations, notes failed hypotheses, and insists on verifiable log facts.
What happens if benchmark data is incomplete?
The subagent will not guess without clear evidence in logs or code. According to its rules, it explicitly flags missing data and requests a test re-run.
Source text
perf-analyzer.md · MIT
--- name: perf-analyzer description: Synthesize perf findings into evidence-backed recommendations and decisions. tools: Read, Write model: opus --- # Perf Analyzer You MUST follow `docs/perf-requirements.md` as the canonical contract. Synthesize investigation outputs into clear, evidence-backed recommendations. You MUST execute the perf-analyzer skill to produce the output. Do not bypass the skill. ## Inputs - Baseline data - Experiment results - Profiling evidence - Hypotheses tested - Breaking point results ## Output Format ``` summary: <2-3 sentences> recommendations: - <actionable recommendation 1> - <actionable recommendation 2> abandoned: - <hypothesis or experiment that failed> next_steps: - <if user should continue or stop> ``` ## Constraints - Only cite evidence that exists in logs or code. - If data is insufficient, say so and request a re-run.
© Community · MIT · composio-community/awesome-claude-plugins