web-performance-debugger CLAUDE.md

Claude Code instructions for jantimon/web-performance-debugger, covering claude.md, what this is, commands, architecture and two execution modes (this is the central design fork).

Instructions file

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 instructions/jantimon/web-performance-debugger/claude-md
Clone the repo
git clone --depth 1 https://github.com/jantimon/web-performance-debugger
Per session 11,255 This file is loaded in full into every session.
When invoked 11,255 The same file — it is already loaded in full.
Security scan C 1 finding. 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 $0.11255 $0.11255
Opus 5 $0.05627 $0.05627
Sonnet 5 $0.02251 $0.02251
Haiku 4.5 $0.01125 $0.01125

Measured yesterday against content hash 252c705f010a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade C, and why

web-performance-debugger CLAUDE.md scanned grade C with 1 finding 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 yesterday.

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.

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

- Per the user's global rule: use `trash`, never `rm -rf`.
CLAUDE.md · 532 lines

How it starts

The opening of the file, as written. The whole thing — 532 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

What this is

wpd (package @jantimon/web-performance-debugger, bins wpd / web-performance-debugger) is a TypeScript CLI that drives Chrome or Firefox via Puppeteer to attribute layout/paint/style/ invalidation work back to source lines, plus CPU sampling (on by default) that attributes self-time to source/package. One user-facing axis picks where it runs: --target chrome|firefox|node. Trust tiers, keep them straight: counts (trace-derived, windowed to the renderer main thread) are exact; slice ms on a --breakdown bar and wall/INP are wall-tier directional (trace base::TimeTicks / Chrome-clamped performance.now()); CPU self-time comes from the profiler's own microsecond clock (not performance.now()), so its ms are a real signal, trustworthy in aggregate (sampling noise ~few %). So it is not a wall-clock benchmark runner, but it is the right tool for comparing JS cost (e.g. SSR renderToString lanes).

selfMs is not "pure JS" on the browser lanes. It is JS plus the synchronous engine work JS triggered: a forced layout lands as self-time on the line that forced it (measured: ~85% of the forced-layout probe's "JS" self-time is reflow). Only --target node (no DOM) measures pure JS. This is a feature — it prices "delete this line" — but do not describe it as pure JS.

Read README.md for the user-facing surface; this file is the internal map; docs/dev/ holds the measured facts behind the non-obvious choices (index) and is the first stop before changing the capture modes, the Gecko converter, or any cross-engine claim.

Commands

npm run build           # tsc -> dist/ (ESM, NodeNext)
npm test                # unit only (pretest builds first); pure functions, no browser
npm run test:e2e        # e2e: drives the real CLI against headless Chrome (record -> query)
npm run lint            # oxlint src test scripts examples   (lint:fix to autofix)
npm run format          # oxfmt --write (format:check to verify; config in .oxfmtrc.json, ~prettier)
npm run knip            # dead exports/files/deps (config knip.json); a fresh dead export fails it
node dist/cli.js <...>  # run the CLI (installed bins: web-performance-debugger, wpd)
npm run changeset       # add a changeset; CI Release workflow versions+publishes on merge to main

Read the full file on GitHub · 532 lines

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. yesterday First seen · 532 lines · 11,255 tokens per session scan C 252c705f010a

Subscribe to this mod's changes

web-performance-debugger CLAUDE.md is an instructions file published in the GitHub repository jantimon/web-performance-debugger (31 stars, last pushed 27d ago), licensed MIT. It adds 11,255 tokens to every session, about $0.0563 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-01.