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 verification-before-completiongit 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/verification-before-completion)<a href="https://agentmods.dev/skills/androoagi/starnet/verification-before-completion"><img src="https://agentmods.dev/badge/skills/androoagi/starnet/verification-before-completion/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/verification-before-completion"><img src="https://agentmods.dev/badge/skills/androoagi/starnet/verification-before-completion.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 | $0.00052 | $0.01545 |
| Opus 5.5 | $0.00021 | $0.00618 |
| Sonnet 5.5 | $0.00010 | $0.00309 |
| Haiku 4.5 | $0.00005 | $0.00154 |
Grade A, and why
verification-before-completion 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 — 155 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Verification Before Completion
Overview
Core principle: evidence before claims, always.
Violating the letter of this rule is violating the spirit of this rule.
The Iron Law
NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE
If you have not run the check in this turn and read what it returned, you cannot claim it passes.
The gate
BEFORE claiming any status or expressing satisfaction:
1. IDENTIFY What check proves this claim?
2. RUN Execute the FULL check (fresh, complete)
3. READ Full output, the exit code, the count of failures
4. VERIFY Does the output confirm the claim?
- If NO: state the actual status, with the evidence
- If YES: state the claim WITH the evidence
5. ONLY THEN make the claim
Skipping any step is misreporting, not verifying.
How to run the check here
- Code (tests, build, lint, type-check):
verify.runwithcmdset to the project's own check (for examplenpm test); it returns a clear PASS or FAIL. Useshell.execfor any other command, and read the combined output AND the exit code it returns. A long suite may need a largertimeoutMs; a run that timed out proved nothing. - A file you wrote or changed: read it back with
fs.read(or find the changed lines withfs.search) and compare it with what was asked. "I calledfs.write" is not evidence that the content is right. - A web or API change: send the real request with
web_requestand read the status and body, or open the page withbrowser.navigateand read it withbrowser.get_textorbrowser.screenshot. - A message, post or booking: read the tool's response. Success is what the response says, not the fact that you made the call.
- A deliverable for the Commander: open the finished file and check it against each requirement before naming it with
deliverable_note. - No way to check (no WORKBENCH placed, or a scheduled run, which has no shell and cannot ask questions): do not guess. Report the work as unverified and name the exact check that still has to be run.
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 · 155 lines · 52 tokens per session scan A 63bea902b43d
verification-before-completion is a skill published in the GitHub repository androoAGI/starnet (1,468 stars, last pushed 3d ago), licensed MIT. It adds 52 tokens to every session and 1,545 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
Check work (verify against criteria)
Verify an implementation against acceptance criteria with a reviewer and a tester.
PR babysit (watch & drive to green)
Watch a pull request and delegate fixes until CI is green and reviews are addressed. Requires the GitHub MCP server or gh.
Write tests
Add focused, meaningful tests for code — happy path plus the edge cases that matter.
build-test
Run the project's build / typecheck / lint / test commands and emit the build.passing + tests.passing signals devloop convergence reads.
verification-before-completion
Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always.
developing-openrig
Use when changing OpenRig's own source in a clone of the openrig repository: finding which package or file owns a behaviour, choosing which tests to run, testing a change without disturbing the OpenRig daemon your own session runs on, judging whether a diff touches a high-risk area, or preparing a pull request. Not…