Agent Sh Perf
agent-sh-perf is a Claude Code plugin for the Ship phase that guides baselines, profiling, and evidence-backed performance decisions.
by agent-sh · github.com/agent-sh/perf
Run a disciplined performance investigation with baselines, profiling, and hypothesis-driven fixes before you ship slow code.
Add it to Claude Code
Install the plugin in Claude Code. One command, paste-ready.
/plugin install agent-sh-perf@agent-sh/perfBuilt to be called by your agent
Skillselion is itself an MCP server. Your agent can pull this entry and a paste-ready install config straight from the API - no copy-paste.
Retrieve this entry with skillselion.get_details("plugin:agent-sh/perf") and the paste-ready config with skillselion.get_install_config("plugin:agent-sh/perf").
What it does
agent-sh-perf packages a performance investigation methodology for builders who ship web apps, APIs, and CLIs and cannot afford vague advice like optimize the database. The plugin emphasizes baselines first: capture numbers, form hypotheses, profile the hot paths, then change one variable at a time with evidence that justifies each decision. That rhythm fits indie teams without a dedicated performance engineer—especially when a feature branch suddenly doubles p95 latency or memory spikes under agent-driven load tests. It spans Ship when you are clearing launch checklists and Operate when production metrics drift, but the catalog shelf stays Ship/perf because that is where most solo builders first feel pain. Intermediate complexity: you should be comfortable reading profiler output or simple benchmarks. Pair it with your existing test harness; it is a workflow skill, not a hosted APM product.
Highlights
- Rigorous performance investigation workflow with explicit baselines
- Profiling-guided steps instead of guess-and-check optimizations
- Hypothesis tracking with evidence-backed accept or reject decisions
- Suited to regressions caught in CI or staging before production ship
- agent-sh perf plugin bundle (single plugin entry in catalog)
Why builders use it
Performance regressions stall launches because optimizations are anecdotal instead of measured against clear baselines and profiles.
You follow a repeatable investigation path from baseline to profile to validated fix, with documented evidence for each performance decision.
At a glance
- Type - Plugin in Development Tools.
- Adoption - 0 installs, 2 stars, 0 votes.
FAQ
Who is agent-sh-perf for?
Developers shipping backends or full-stack products who want agent-guided performance investigations with baselines and profiling, not random tuning tips.
When should I use agent-sh-perf?
Use it when latency, throughput, or memory regressed and you need hypotheses tested against measurements before merging or releasing.
How do I add agent-sh-perf to my agent?
Add the agent-sh/perf Claude Code plugin from Skillselion, ensure your project can run benchmarks or profilers locally, and start the perf workflow from the plugin commands.
Comments
Share how you use agent-sh-perf, gotchas, or tips for other indie builders.
No comments yet - be the first to share how you use it.