StarNet is a local-first desktop harness for creating AI agents and arranging them in a pixel-art space station whose layout represents their capabilities, permissions, and workflow. It lets people run real model calls and tools through multiple bounded agents, using their own provider keys or accounts and optionally messaging the station through chat services. The catalogue entries are skills that extend or support work performed by StarNet agents.
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 androoAGI/starnet --skill studio-videogit clone --depth 1 https://github.com/androoAGI/starnetWrote 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/androoagi/starnet/studio-video)<a href="https://agentmods.dev/skills/androoagi/starnet/studio-video"><img src="https://agentmods.dev/badge/skills/androoagi/starnet/studio-video/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/androoagi/starnet/studio-video"><img src="https://agentmods.dev/badge/skills/androoagi/starnet/studio-video.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 | $0.00044 | $0.00949 |
| Opus 5.5 | $0.00018 | $0.00380 |
| Sonnet 5.5 | $0.00009 | $0.00190 |
| Haiku 4.5 | $0.00004 | $0.00095 |
Grade A, and why
studio-video 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 12d 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 — 38 lines — stays where its author put it; the contents beside it link to each section on GitHub.
The studio cuts a video from parts it makes itself: voiceover takes from voice_generate, stills from image_generate, and an edit decision list from resolve_timeline_file. The narration is the spine — pictures are timed to the voice, never the other way round.
Method
- Write the script as numbered beats. One beat = one idea, one or two sentences of narration, one picture. Note for each beat what is on screen. Spoken narration runs roughly 150 words a minute; size the script to the target length.
- Get the script approved before rendering anything. Voice and images cost money; a rewrite after rendering pays twice.
- Record one take per beat (voice_generate). Plain prose, no markup. Name each take with
pathso the order is obvious:audio/v01-intro.mp3,audio/v02-problem.mp3. Use the samevoiceandstylefor every take. Text over 4,000 characters is refused, so split long beats — never truncate. - Make the stills (image_generate). One shape for all:
aspect_ratio: "16:9"for widescreen,"9:16"for phone video. Write one shared style line and repeat it word for word in every prompt; only the subject changes. Name stills to match the takes:images/v01-intro.png. For a frame that must show readable words, use the premium model the tool description names. - Look before you cut (image_analyze). Check each still for garbled text, the wrong subject, or style drift. Regenerate failures; do not cut around them.
- Measure the takes. Each still holds for the length of its take. With a WORKBENCH, read real durations with ffprobe through shell.exec; without one, estimate from word count and let the timeline tool check you — it reads durations with ffprobe when available and notes any out-point past the end of a file.
- Build the timeline (resolve_timeline_file). Main track (lane 0): the stills in beat order, each with
in: 0andout= its hold in seconds; they play back to back. Voiceover: each take on lane -1 withat= the second its beat starts. Add a marker per beat, named for the beat. Setfps(24 unless asked) and thewidth/heightthat match the stills. The file lands atedits/<name>.fcpxmlunless you passpath. - Read the tool's notes. A "past the media end" or "not probed" note means the cut is wrong or unverified — fix the numbers and rebuild.
- Open it. Call resolve_status. If live control is YES (Resolve Studio), use resolve_control action=import_timeline with the file; render only when asked (action=render writes under
renders/, then action=render_status). If live control is NO, tell the Commander: in Resolve, File > Import > Timeline, and pick the file. - Name the deliverable (deliverable_note) with the timeline file as
main.
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.
- 12d ago First seen · 38 lines · 44 tokens per session scan A 5f0ba0ef94e2
studio-video is a skill published in the GitHub repository androoAGI/starnet (1,468 stars, last pushed 3d ago), licensed MIT. It adds 44 tokens to every session and 949 once invoked, about $0.0002 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
Review a GitHub PR (via gh)
Review a specific GitHub pull request with gh — fetch the diff, fan out reviewers, consolidate, and optionally post the review. Requires the gh CLI or the GitHub MCP server.
Implement (multi-agent loop)
Orchestrate an implement -> review -> fix loop with subagents until reviewers sign off.
Best of N (parallel attempts)
Delegate N parallel subagents on the same task, then pick the best result.
Check work (verify against criteria)
Verify an implementation against acceptance criteria with a reviewer and a tester.
Commit (clean, conventional)
Stage the right changes and write a clear, conventional commit message.
Design doc (write -> review loop)
Draft a design document and iterate writer/reviewer subagents until consensus.