By alastor0325
Firefox bug investigation toolkit — bug-start hub plus profile/log analysis, spec checking, code navigation, investigation write-up, and patch review. Component-agnostic mechanics; domain knowledge runs deepest for A/V/media. Shared wiki is an optional accelerator.
Single-dimension Firefox patch reviewer. Spawned only by the `/review` skill orchestrator — never invoked directly by a user. Reviews one patch through exactly one lens (spec, security, threading, IPC, error-handling, api-usage, code-quality, or tests), or verifies/refutes one finding, and returns structured JSON only.
Use this agent to understand Firefox Gecko codebase architecture, trace execution flows, locate symbol definitions, and navigate complex subsystems (especially dom/media). Read-only specialist for code orientation, not implementation. **Trigger patterns - Use when the user asks:** - "How is X used in Gecko?" (e.g., "how is PlayReady EME used") - "Explain how X works" (e.g., "explain seamless looping") - "How does X integrate with Y?" (e.g., "how does WMF plug into media pipeline") - "Where is X implemented?" - "What owns/controls X?" - "Trace the flow from A to B" - Questions about architecture, threading, object lifetimes, or subsystem integration - Understanding DRM/EME, media playback, graphics, IPC, or other complex subsystems
Analyze a Firefox Profiler capture (share.firefox.dev or profiler.firefox.com/public links) for performance and media-pipeline issues, producing a structured findings report consumed by bug-start and triage. Triggers on "analyze profile", "analyze this profile", "/analyze-profile", or a Firefox Profiler URL.
Quick start workflow for investigating or implementing a Firefox bug
Analyze a Firefox log file for problems. Specializes in media/EME/CDM issues (hardware context reset, process crashes, HRESULT errors, PlayReady/CDM failures). Extensible to other Firefox subsystems.
The single approval gate for downloading any external file (bug attachment, media sample, test fixture, profile, screen recording, etc.). Never auto-download — surface a Yes/No AskUserQuestion, and on Yes fetch into the one shared transient folder ~/.fx-bug-toolkit/download-cache/. Invoke this whenever any skill or flow (/triage, /bug-start, the Process-queue drain, …) needs a file pulled to disk.
One-time setup for the fx-bug-toolkit plugin — walks two checklists (environment/paths, then dependencies) one item at a time, ensuring every REQUIRED item is set/installed and confirming OPTIONAL ones can fall back to a default. Creates the data dirs and installs missing CLIs with your confirmation. Triggers on "init fx-bug-toolkit", "set up the toolkit", "fx-bug-toolkit setup".
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A Claude Code plugin that helps you investigate Firefox bugs — fast.
Point it at a bug and it pulls the Bugzilla report, searches the Gecko codebase,
reads profiles and logs, checks the spec, and writes up a clean, link-rich
investigation file for you. One command does the whole thing: bug-start.
It works for any Firefox component — the mechanics aren't tied to a particular area. The bundled know-how currently runs deepest for media / A/V (that's where it grew up), and it gets smarter for other areas as people add knowledge to the shared wiki.
This toolkit investigates and diagnoses — it doesn't write or land patches. Think of it as the "understand the bug" half of your workflow.
New here? The Getting Started tutorial is the friendliest way in — install, your first investigation, and the viewer, with screenshots. (It's an interactive page; open it in a browser.) This README is the quick reference.
About the commands: you can type the short name —
/init,/bug-start, etc. — and Claude Code's command picker shows which plugin each one comes from, so you can pick thefx-bug-toolkitone. The fully-qualified form (/fx-bug-toolkit:init) always works too, if you'd rather be explicit (handy when a personal skill or another plugin shares the same name).
1. Install the plugin (no clone needed):
claude plugin marketplace add alastor0325/fx-bug-toolkit
claude plugin install fx-bug-toolkit@fx-bug-toolkit
Prefer a local copy to read or hack on? Clone first, then point the marketplace at the folder:
git clone https://github.com/alastor0325/fx-bug-toolkit.git ~/fx-bug-toolkit
claude plugin marketplace add ~/fx-bug-toolkit
claude plugin install fx-bug-toolkit@fx-bug-toolkit
2. Restart Claude Code so the plugin loads. (Plugins are picked up at startup.)
3. Run the setup — it checks your machine and offers to install anything missing:
/init
When init says "Setup complete," you're good to go.
/bug-start <bug-id> runs the whole investigation; /open-investigation opens
the viewer to search and re-read past ones. The Getting Started tutorial
walks through it with examples and screenshots — start there.
Tip: type / and start typing a name to see these in the picker, with the
source plugin shown next to each.
| Command | What it does |
|---|---|
/init | One-time setup — checks your env + dependencies, offers to install what's missing |
/bug-start <bug-id> | The hub. Investigate a Firefox bug end to end and write the investigation file |
/analyze-profile <url> | Analyze a Firefox Profiler capture |
/check-log <path> | Diagnose a Firefox log (great for media/EME/CDM crashes) |
/open-investigation | Open a local web viewer to search & read all your past investigations |
/triage | Firefox A/V weekly bug triage — sweep watched + recent bugs, write reviewable drafts |
/open-triage | Open the triage dashboard (web UI over the drafts); lazily installs it on first use |
/review <D-rev|local|diff> | AI patch review — routes the patch to per-aspect reviewers (always spec, security/memory-safety, threading, code quality, tests; plus IPC, error-handling, API usage when relevant), adversarially verifies the serious findings, and writes a structured review doc |
/open-review [repo] | Open Revue, the local web UI for reviewing a repo's patches by hand. Pass a repo/worktree path, or omit it to reuse your remembered repo (Revue's revue init default in ~/.revue/config.json); lazily installs Revue on first use |
/open-team [team] | Open the firefox-review-stats team dashboards (review-load distribution + a digest of what each component shipped). Opens the live site — landing page by default, or playback/webrtc/gfx when named. No install |
/update | Update the plugin + its CLI dependencies to the latest |
npx claudepluginhub alastor0325/fx-bug-toolkit --plugin fx-bug-toolkitFirefox Knowledge Wiki — persistent knowledge base with automatic ingest, lookup, lint, and monitoring hooks
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
AI-powered wiki generator for code repositories. Generates comprehensive, Mermaid-rich documentation with dark-mode VitePress sites, onboarding guides, deep research, and source citations. Inspired by OpenDeepWiki and deepwiki-open.
Claude + Obsidian knowledge companion. Sets up a persistent, compounding wiki vault (Karpathy's LLM Wiki pattern). v1.7 "Compound Vault" + v1.8 methodology modes close 5 of 5 priority gaps from the May 2026 compass artifact. Ships: substrate alignment with kepano/obsidian-skills, default Obsidian CLI transport, hybrid retrieval (contextual prefix + BM25 + cosine rerank per Anthropic's Sept 2024 research), per-file advisory locking for multi-writer safety, pre-commit verifier agent, AND methodology modes (LYT / PARA / Zettelkasten / Generic) for first-class organizational support no other Claude+Obsidian competitor offers. v1.7.x audit closure: every BLOCKER + HIGH + MEDIUM + LOW finding from the v1.7.0 audit is CLOSED or DEFERRED-with-rationale. Optional DragonScale Memory extension (log folds, deterministic addresses, semantic tiling lint, boundary-first autoresearch).
Complete AI coding workflow system. Self-correcting memory + persistent FTS5-indexed research wikis + auto-research loop + multi-LLM council on a single SQLite store. 33 skills, 8 agents, 22 commands, 37 hook scripts across 24 events. Cross-agent via SkillKit.
Make your AI agent code with your project's architecture, rules, and decisions.