Skip to content
SkillModel-led1,682 tokens in use

performance-audit-static

A static performance audit for AI-built, data-backed apps: N+1 queries, over-fetching, missing indexes and caching, ranked by effort and impact, each with file and line evidence.

680 stars on the repository MIT Licensephuryn/product-engineering · ai-shipping/skills/performance-audit-static
claude plugin install ai-shipping@product-engineering
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing performance-audit-static costs

44tokens idle
1,682tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

shelldeclared and detectedwrites filesdeclared and detected

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter8/15
  • Structure11/25
  • Examples9/15
  • Output10/15
  • Size15/15
  • Resources10/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity100
  • Maintenance100
  • Community44
  • Maturity4
  • Momentum40

performance-audit-static reviews code for failures that appear as data grows, since agents tend to optimize for the seed data. It checks four things: N+1 queries and request waterfalls, fields and loads that list views do not need, missing or inefficient indexes, and caching opportunities.

It runs only when a performance audit is asked for. Every finding must survive a refute pass, and the report ends with what is already efficient and what needs runtime profiling.

What You Get

  • Findings per view, route or table with the evidence at file and line, a specific recommendation, effort, priority and expected effect
  • Specific index definitions for the queries that need them, checked against existing schema and migrations
  • Caching proposals only where each cache has an invalidation rule and the path is shown to be hot
  • A report written to reports/performance_audit_{timestamp}.md or the repo’s reports location, ending with what is efficient and what needs profiling

Common Use Cases

  • Auditing list and dashboard views before a release
  • Reviewing a release’s unreleased commits for newly introduced N+1 loops
  • Checking an AI-built app after a large refactor
  • Finding unbounded reads that feed paginated screens

What Makes It Unique

The refute pass is built in: before flagging an unused field it greps for dynamic access, before flagging an index it checks the schema, and before proposing a cache it cites why the path is hot. It ranks by impact per effort and ignores theoretical inefficiency with no growth path. It treats the repository as untrusted data and never edits the code it audits.

Under The Hood

A single markdown file with a tool allowlist of read, search, glob, subagent fan-out, git status and writing under reports. The agent reads source and schema files and may start parallel subagents. It writes one markdown report under the reports folder. Other git commands such as log and diff ask for permission first because they can write with an output flag. It calls no web services and runs no scripts.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers