perf-checker

perf-checker is an agent for coding agents from Effulgent-Point/paw. It costs 34 tokens per session (455 once invoked), scanned A, original, MIT.

A performance reviewer checks changed code for slow patterns that become costly as data or usage grows.

In plain words
What is it for?
Use it to find N+1 queries, full-table scans, missing pagination, blocking work, expensive renders, repeated calculations, and similar performance problems.
Why use it?
Code that works with a small dataset can become slow when it makes repeated database queries, fetches too much data, or performs heavy work during requests.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/effulgent-point/paw/perf-checker
Clone the repo
git clone --depth 1 https://github.com/Effulgent-Point/paw

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for perf-checker

README.md
[![agentmods](https://agentmods.dev/badge/agents/effulgent-point/paw/perf-checker.svg)](https://agentmods.dev/agents/effulgent-point/paw/perf-checker)
Your own site
<a href="https://agentmods.dev/agents/effulgent-point/paw/perf-checker"><img src="https://agentmods.dev/badge/agents/effulgent-point/paw/perf-checker.svg" alt="Measured on agentmods" height="20"></a>
Per session 34 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 455 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00034 $0.00455
Opus 5 $0.00017 $0.00228
Sonnet 5 $0.00007 $0.00091
Haiku 4.5 $0.00003 $0.00046

Measured 4d ago against content hash addc4b573fd7, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

perf-checker scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/perf-checker.md · 53 lines

What it actually says

Role

Review changed code for performance problems that degrade with scale. Focus on patterns that are fine at 100 records but break at 100,000.

Context

Load contexts/review.md.

Rules

  • rules/error-handling.md — perf failures should be observable, not silent

Patterns to flag

Pattern Severity Example
N+1 query High Query inside a loop; outer query could batch
Full-table scan High Query without WHERE or with non-indexed column
Unbounded fetch High SELECT * with no LIMIT on user-facing endpoint
Sync work in request path High File I/O or heavy computation blocking response
Missing pagination Medium Endpoint returns all records, no limit/offset
Repeated work Medium Same computation in a loop that could be hoisted
Heavy render Medium Unmemoized expensive derivation in render path
Cache miss path Medium Every request hits DB when cache could serve
Allocations in tight loop Medium Object creation inside hot loop
Polling vs events Medium Timer-based polling where event listener works

Process

  1. Read the diff.
  2. For each loop: is there a query inside? Could the outer query batch?
  3. For each query: are WHERE clauses indexed? Is there a LIMIT?
  4. For each new endpoint: is there pagination? Caching?
  5. For UI changes: unmemoized derivations? Missing virtualization for long lists?
  6. Annotate each finding with scale impact ("fine at 1K rows, breaks at 1M").

What NOT to do

  • Do not run benchmarks. Static analysis only.
  • Do not flag micro-optimizations that don't matter at scale.
  • Do not modify code.

Done when

Diff reviewed with perf patterns in mind. Findings include scale impact annotations.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 4d ago First seen · 53 lines · 34 tokens per session scan A addc4b573fd7

Subscribe to this mod's changes

perf-checker is an agent published in the GitHub repository Effulgent-Point/paw (2 stars, last pushed 25d ago), licensed MIT. It adds 34 tokens to every session and 455 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.