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 skills add n2ns/antigravity-panel --skill devhost-debuggit clone --depth 1 https://github.com/n2ns/antigravity-panelWrote 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.
[](https://agentmods.dev/skills/n2ns/antigravity-panel/devhost-debug)<a href="https://agentmods.dev/skills/n2ns/antigravity-panel/devhost-debug"><img src="https://agentmods.dev/badge/skills/n2ns/antigravity-panel/devhost-debug/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/n2ns/antigravity-panel/devhost-debug"><img src="https://agentmods.dev/badge/skills/n2ns/antigravity-panel/devhost-debug.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00084 | $0.01028 |
| Opus 5.5 | $0.00034 | $0.00411 |
| Sonnet 5.5 | $0.00017 | $0.00206 |
| Haiku 4.5 | $0.00008 | $0.00103 |
Grade A, and why
devhost-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 6d 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.
How it starts
The opening of the file, as written. The whole thing — 46 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Extension Development Host debugging
The F5 Extension Development Host is a real Antigravity IDE window. Its extension host runs in WSL on this machine with an inspector port, it has its own live Language Server, and its renderer is reachable over CDP when the IDE was started with --remote-debugging-port=9222. devhost.mjs finds all of this by itself; the inspector port changes on every launch.
node .claude/skills/devhost-debug/devhost.mjs <command> [args]
| Command | What it does |
|---|---|
status |
Extension host PID and inspector port, Language Server PID, log file, dev host page, sidebar version and Auto-Accept state |
log [--lines N] [--grep re] |
Tail of the dev host "Antigravity Panel" output log |
shot [out.png] |
Screenshot of the dev host window; pass a path in your scratchpad directory, then view it with Read |
webview '<expr>' |
Evaluate in the sidebar webview; doc is the sidebar document, e.g. doc.querySelector('.auto-accept-status')?.innerText |
page '<expr>' |
Evaluate in the dev host workbench page (Agent panel lives here as .antigravity-agent-side-panel) |
host '<expr>' |
Evaluate in the extension host global scope (process, process.getBuiltinModule); module internals are not reachable here |
logpoint '<source regex>' '<expr>' [--seconds N] [--max N] |
Non-pausing logpoint in dist/extension.js: records <expr> each time the matched location runs, then removes itself |
dryrun [--terminal] |
Auto-Accept scan built from src, with the click and DOM marker stripped, run on the dev host page; prints { panel, events } |
Workflow
npm run build(or keepnpm run watchrunning, see logpoints).- Ask the user to reload the Extension Development Host window. Do not reload it yourself.
status, thenshot/webview/logto check the result;logpointfor extension host state;dryrunfor Auto-Accept scan changes.- Report what was observed in the dev host separately from unit test results, and say which states could not be produced (for example no pending Agent action).
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 6d ago First seen · 46 lines · 84 tokens per session scan A 25ec45399d38
devhost-debug is a skill published in the GitHub repository n2ns/antigravity-panel (719 stars, last pushed 6d ago), licensed Apache-2.0. It adds 84 tokens to every session and 1,028 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-10-04.
Other skills, from other repositories
graph-it-live
Analyze code dependencies, call graphs, and architecture using the graph-it CLI. Use this skill whenever the user wants to understand how their code is connected — even if they don't say "dependency graph" explicitly. Trigger for: "what calls this function", "who uses this class", "is it safe to delete X", "what…
dead-code-hunter
Scan the dependency graph for orphan nodes (uncalled symbols and unreferenced files) and produce a safe, ranked deletion plan. Use this skill whenever the user asks about dead code, unused code, orphan symbols, unreferenced files, code cleanup, or code hygiene — even if they don't say "dead code" explicitly. Also…
debugging-and-error-recovery
A systematic debugging procedure for finding the underlying cause of failed tests, broken builds, bugs, and unexpected behavior. It emphasizes reproducing the problem, preserving evidence, fixing the cause, and checking the fix.
second-opinion
Get an independent review of the current change from a different model family (GPT via Codex, Copilot or Cursor; Gemini via Antigravity) through agent-intern, then check each finding against the code and reconcile it with your own view. Use before merging a risky diff, when stuck on a bug, or when the user asks for a…
performance-analysis
Comprehensive performance analysis, bottleneck detection, and optimization recommendations for Claude Flow swarms.
debugging-and-error-recovery
Guides systematic root-cause debugging with hard rules against guess-fixes and symptom suppression. Use when tests fail, builds break, behavior doesn't match expectations, or you encounter any unexpected error. Triggers on "this is broken", "tests are failing", "why doesn't this work", or any error output.