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 address-pr-commentsgit 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/address-pr-comments)<a href="https://agentmods.dev/skills/androoagi/starnet/address-pr-comments"><img src="https://agentmods.dev/badge/skills/androoagi/starnet/address-pr-comments/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/address-pr-comments"><img src="https://agentmods.dev/badge/skills/androoagi/starnet/address-pr-comments.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 | $0.00046 | $0.01100 |
| Opus 5.5 | $0.00018 | $0.00440 |
| Sonnet 5.5 | $0.00009 | $0.00220 |
| Haiku 4.5 | $0.00005 | $0.00110 |
Grade A, and why
address-pr-comments 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 review comment is a claim about the code, and the reviewer may be right, wrong or unclear. Check each one against the code before you act, and leave no thread without an honest answer. The bar: every open thread ends as a change with its commit, a reasoned reply, or a question.
Method
- Pick the road to GitHub. Run
gh --versionandgh auth status(shell.exec). Both succeed: usegh. Otherwise look for the connected GitHub account's tools (tool.search "github") and use the names it returns, never guessed ones. Neither: stop and tell the Commander to install and sign in to the GitHub CLI from its official site, or connect GitHub in ABILITIES. - Get on the pull request's branch.
git statusfirst — uncommitted work that is not yours means stop and ask. Thengh pr checkout <number>and confirm the branch and its latest commit match the pull request. - List every open thread. Inline comments:
gh api repos/{owner}/{repo}/pulls/<number>/comments --paginate(each hasid,path,line,in_reply_to_id). Review summaries and general comments:gh pr view <number> --comments. Which threads are already resolved is only in the thread query inreferences/review-threads.md; if you cannot read that, treat every thread as open and say so. Put the count on your todo list — that is the checklist. - Group by file and read each thread to its end. A later reply may withdraw or change the ask. Mark threads that make the same point in several places so one change answers them together.
- Verify the point before acting. fs.read the file as it is now (
numbered: true) — the code may have moved since the comment was written. Trace the path the reviewer describes, or run the case. Then sort the thread: agree, disagree or unclear. - Agree: change it. The smallest change that answers the point (fs.edit / fs.patch), committed by itself or with its tightly related threads so the table can name the commit. Draft the reply: what changed and where.
- Disagree: answer with evidence. No change. Draft a courteous reply giving the reason and the proof — the line, the test output, the documented behaviour. The reviewer and the Commander decide from there.
- Unclear: ask. Draft one specific question. Do not guess at a change.
- Run the tests. The project's own test command plus the lint and type checks its CI runs (verify.run / shell.exec). A check that was already red before your commits is a baseline: note it, do not chase it.
- Show and hold. Give the Commander the table, the diff and every drafted reply. With the go-ahead, push, then post each reply on its own thread (the reply command is in the reference file, or the connected account's reply tool), naming the commit. Re-read the thread to confirm each reply landed.
What ships with it
2 files 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 · 46 tokens per session scan A ab72bc1c9d90
address-pr-comments is a skill published in the GitHub repository androoAGI/starnet (1,468 stars, last pushed 3d ago), licensed MIT. It adds 46 tokens to every session and 1,100 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.
Design doc (write -> review loop)
Draft a design document and iterate writer/reviewer subagents until consensus.
Review changes (multi-agent)
Delegate read-only reviewers over the local changes/branch and consolidate findings.
Refactor (behavior-preserving)
Improve code structure in small, safe steps without changing behavior.
diff-review
Render the patch-edit run's accumulated changes as a reviewable diff, surface it through a GenUI choice surface, and persist the user's accept / reject decision into the artifact manifest.