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.
npx agentmods add commands/extensiondev/mcp/extension-debuggit clone --depth 1 https://github.com/extensiondev/mcpWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00013 | $0.00773 |
| Opus 5 | $0.00006 | $0.00387 |
| Sonnet 5 | $0.00003 | $0.00155 |
| Haiku 4.5 | $0.00001 | $0.00077 |
Grade A, and why
extension-debug 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 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.
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.
How it starts
The opening of the file, as written. The whole thing — 53 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Debug the currently running extension dev session. The user said: $ARGUMENTS
Steps
-
Check for a running dev session
- Look for
dist/extension-js/chrome/ready.jsonin the project root - If MCP tool
extension_waitis available, use it with a short timeout (3s) to check - If no session: tell the user to start one with
/extension devornpm run dev
- Look for
-
Inspect the live state If MCP tool
extension_inspectis available:- Pass
include: ["html", "summary", "meta", "console", "extension_roots"] - If the user provided a URL in
$ARGUMENTS, pass it asurl - If the user provided CSS selectors (strings starting with
#,., or[), pass them asprobe - Report the results in a structured way
If MCP is not available:
- If the session was started with
--allow-control, suggest the CLI equivalent:npx extension inspect --tab <id> --include summary,html --with-console 20 - Otherwise read
dist/extension-js/chrome/ready.jsonto get the CDP port and suggest Chrome DevTools inspection
- Pass
-
Exercise event handlers (when the bug is in an action / command / shortcut) If the session was started with
--allow-control, fire the events a user would, without clicking:extension_openwithsurface: "action", triggers the toolbar action (opens its popup, or replayschrome.action.onClicked).extension_openwithsurface: "command"andname: "<cmd>", replays achrome.commands.onCommandkeyboard shortcut.- Then re-inspect or read
extension_logsto see what the handler did.
Caveat: replay carries no user gesture, so
activeTabis NOT granted. If the handler depends onactiveTab(e.g.chrome.scripting.executeScripton the active tab,captureVisibleTab), the result includesgesture: falseand awarning, and behavior differs from a real click. For a genuine-gesture click (activeTab granted), use chrome-devtools-mcp'strigger_extension_action(Chromium only) alongside this server.To see what else is loaded in the browser (Chromium):
extension_list_extensions. -
Diagnose common issues Based on what you find, check for:
- "It didn't load": Check extension root count. If 0, content scripts may not be injecting. Check manifest
content_scriptsmatches patterns and the target URL. - Console errors: Report top errors. Common ones:
Uncaught TypeError, likely a missing import or wrong module formatContent Security Policy, extension CSP blocking inline scriptschrome.runtime.lastError, API permission missing
- Wrong page: Check if the content script
matchespattern in manifest covers the target URL - Shadow DOM empty: Extension root exists but shadow content is empty, likely a CSS or framework mounting issue
- "It didn't load": Check extension root count. If 0, content scripts may not be injecting. Check manifest
-
Suggest fixes based on the diagnosis
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.
- yesterday First seen · 53 lines · 13 tokens per session scan A f20e5900c6b7
extension-debug is a command published in the GitHub repository extensiondev/mcp (1 stars, last pushed 2d ago), licensed Apache-2.0. It adds 13 tokens to every session and 773 once invoked, about $0.0001 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.
Other commands, from other repositories
agentganggang
Use MultiAiPanel when you want Claude Code to compare one prompt across multiple supported AI chat tabs through the local MultiAiPanel MCP surface.
template
Manage issue templates for streamlined issue creation.
design-review
Workflow recipe — review a design end-to-end, ending in measured numbers rather than adjectives, by chaining 4 skills.
setup-pm-skills
Onboard a new user — find out what they do, recommend the right bundles & top skills, and set up a project CONTEXT.md so every skill is tailored to them.
security-review
CWE 기반 보안 검토 + STRIDE 위협 모델링 (v6 - effort:max 강제).
release-harn
Run the tag-first Harn release workflow.