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/yuting0624/antigravity-for-claude-code/mediagit clone --depth 1 https://github.com/yuting0624/antigravity-for-claude-codeWrote 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/commands/yuting0624/antigravity-for-claude-code/media)<a href="https://agentmods.dev/commands/yuting0624/antigravity-for-claude-code/media"><img src="https://agentmods.dev/badge/commands/yuting0624/antigravity-for-claude-code/media.svg" alt="Measured on agentmods" 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 | $0.00038 | $0.00715 |
| Opus 5 | $0.00019 | $0.00358 |
| Sonnet 5 | $0.00008 | $0.00143 |
| Haiku 4.5 | $0.00004 | $0.00072 |
Grade A, and why
media 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 5d 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.
What it actually says
Claude Code can't hear audio or watch video, and doing it locally means an ffmpeg + speech-model stack. Gemini is natively multimodal — so delegate the perception to agy and keep the judgment here.
Input: $ARGUMENTS
Do this:
-
Resolve the file. First arg is the path; anything else (that isn't a flag) is the focus/question. If no file was given, ask which one (AskUserQuestion) — don't guess.
-
Delegate (the engine handles format pre-flight, the digest contract, and writes the full transcript to a file):
agy-media <file> [focus] [--convert] [--tier pro|flash] [--timeout 20m] [--out <path>]- Default tier is
pro(better timestamps/diarization);--tier flashis fine for short/simple clips. - Long media needs headroom: raise
--timeout(e.g.20m) for anything over a few minutes. If it still times out (exit 12), split the file into ~30-min chunks and run them separately. - Exit 5 = unsupported format. agy mishandles
.m4a/.aiff(common for voice memos) even though Gemini accepts them. The engine prints the exact conversion command — re-run with--convertto have it converted automatically (macOSafconvert, elseffmpeg).
- Default tier is
-
Ingest ONLY the digest it prints (summary · timestamped outline · key points · quotes · action items · visuals · uncertainty notes). Do not read the whole transcript file into context — a 1-hour recording is ~10k words and re-reading it every turn is exactly the cost blow-up this plugin exists to avoid. The transcript file is there so you can grep/read slices on demand.
-
Verify before you rely on it (transcription is not ground truth):
- Check the transcript file actually exists (the engine warns if it doesn't).
- The digest flags unclear audio and uncertain names/numbers — treat those as
unverified. For any load-bearing figure, name, or quote, read that timestamp's
slice from the transcript file (e.g.
grep -n "\[12:3" <transcript>) rather than trusting the summary. - Say plainly which claims you confirmed and which are still the model's guess.
-
Report: the digest's substance (not a re-paste), where the transcript lives, and what you verified. If the user asked a specific question, answer it with
[mm:ss]citations so they can jump to the source.
Good uses: meeting/interview notes, a screencast or demo video, a voice memo, a conference talk, a UI walkthrough, an architecture diagram or screenshot.
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.
- 5d ago First seen · 52 lines · 38 tokens per session scan A 93bd8eb754f7
media is a command published in the GitHub repository yuting0624/antigravity-for-claude-code (309 stars, last pushed yesterday), licensed MIT. It adds 38 tokens to every session and 715 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 2026-08-30.
Other commands, from other repositories
agy
Delegate any task to agy (Google Antigravity CLI) — auto-selects model by task type.
ask
Query multiple AI agents (Gemini, OpenAI, Grok, Perplexity, Kimi, any model OpenRouter routes to, and a local ollama model) for diverse perspectives on architecture decisions, technology choices, debugging dead-ends, and security tradeoffs. Use this whenever the user names the council directly, whatever the topic …
result
Fetch, list, or cancel background council jobs started with --async.
setup-video-vision
Interactive setup wizard for claude-video-vision — configure backend, whisper, frames, and verify dependencies.
notebook
Local NotebookLM over a FOLDER of documents using Antigravity (agy). Sweeps each document (PDF with text, scanned PDF, image, docx) into an objective-driven Markdown summary, then builds a relevance INDEX and a cited master summary. Incremental cache (re-runs only re-summarize changed docs / changed objective) and…
deep-research
Deep, multi-source, fact-checked web research with agy — reach for it when a decision or design depends on getting it right and a single-shot answer is not enough (architecture / tool / vendor choices, thorough landscape scans, anything you will act on). Builds an evidence matrix + a plan you approve, then agy browses…