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 link-and-video-summarygit 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/link-and-video-summary)<a href="https://agentmods.dev/skills/androoagi/starnet/link-and-video-summary"><img src="https://agentmods.dev/badge/skills/androoagi/starnet/link-and-video-summary/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/link-and-video-summary"><img src="https://agentmods.dev/badge/skills/androoagi/starnet/link-and-video-summary.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 | $0.00058 | $0.01041 |
| Opus 5.5 | $0.00023 | $0.00416 |
| Sonnet 5.5 | $0.00012 | $0.00208 |
| Haiku 4.5 | $0.00006 | $0.00104 |
Grade A, and why
link-and-video-summary 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 — 41 lines — stays where its author put it; the contents beside it link to each section on GitHub.
A summary stands in for something the Commander will not read or watch, so it has to be true to the source. Cover only what you actually read, point to where each claim sits, and copy quotes exactly.
Method
- Open the source.
web_fetchthe link. Note the title, author, date and the full length the tool reports. If only an empty shell comes back, usebrowser.navigateandbrowser.get_text. - Read all of it, in chunks. For a long page
web_fetchshows the first part and saves the full text to a file it names. Read that file withfs.readin ranges, start to end, taking notes per chunk: the point, and where it sits (heading, paragraph, page or timestamp). Never summarise from the first chunk alone. - PDFs. A PDF link usually reads with
web_fetch; a PDF file on disk does not read withfs.read(that reads text, Word, Excel and image files): ask for a link or a text export, or use the PDF & Document Extraction skill when the WORKBENCH is placed. If no readable text comes back (a scan, for instance), say so and stop. - Threads. Read in order and keep who said what; a reply is not the author's view. If replies are hidden behind a sign-in, say which part you got.
- Video. There is no tool that listens to audio. Try
web_fetchon the video page first: if a transcript or captions come back, use them. Otherwise open the page withbrowser.navigate, find the transcript control withbrowser.snapshot, open it withbrowser.click, and read it withbrowser.get_text. No transcript: tell the Commander it cannot be summarised. The title and description are not the video; offer them only labelled "from the description only". Auto-made captions: say so, because names and numbers may be wrong. - Write the gist. Three lines: what it is and who made it, the main claim, and why it matters.
- List the key points with locations. Five to nine, each followed by where it came from: section heading, page, paragraph or timestamp.
- Copy the quotes exactly. Two to five short, notable quotes, character for character, with the speaker and the location. Find each one again in the source text before you send; if you cannot, drop it.
- Separate opinion from evidence. Sort the main claims: backed (say by what: data, a document, a named source), asserted (the author's view, nothing shown), and disputed or not checkable. You are sorting, not fact-checking; say you did not verify the claims unless asked to.
- Rewrite on request. Notes (headings and bullets), a thread (numbered posts, each making sense alone), or a short post (one paragraph, source credited and linked). Use your own sentences, not the source's rearranged. For a file:
fs.write, thendeliverable_note.
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 · 41 lines · 58 tokens per session scan A ed2002aa19ac
link-and-video-summary is a skill published in the GitHub repository androoAGI/starnet (1,468 stars, last pushed 3d ago), licensed MIT. It adds 58 tokens to every session and 1,041 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.