Borrowing it
Nothing to install: this file belongs to permit0-ai/permit0. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/permit0-ai/permit0/main/.claude/skills/drive/SKILL.mdgit clone --depth 1 https://github.com/permit0-ai/permit0Wrote 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/permit0-ai/permit0/drive)<a href="https://agentmods.dev/skills/permit0-ai/permit0/drive"><img src="https://agentmods.dev/badge/skills/permit0-ai/permit0/drive/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/permit0-ai/permit0/drive"><img src="https://agentmods.dev/badge/skills/permit0-ai/permit0/drive.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 116 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00081 | $0.01855 |
| Opus 5.5 | $0.00032 | $0.00742 |
| Sonnet 5.5 | $0.00016 | $0.00371 |
| Haiku 4.5 | $0.00008 | $0.00186 |
Grade A, and why
drive 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 2mo 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 — 122 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Drive the Polis pipeline
An interactive cockpit for the Polis pipeline. Run it as a loop: show the state panel →
the user picks an action → apply it via gh → (optionally) watch the run → refresh the panel.
The goal is to drive everything from Claude Code; the user never opens github.com. You only
trigger and report — the actual work runs in GitHub Actions (workflow "Agent Pipeline").
Loop
Repeat until the user says they're done:
- Render the State panel (below).
- Ask the user what to do (AskUserQuestion) — offer the suggested next action first.
- Run the chosen action via
gh. Confirm before any action that triggers a cloud run or merges. - If it triggered a run and the user wants, watch it; otherwise refresh the panel.
Step 0 — Preflight
command -v gh— if missing, tell the user to install it and stop.gh auth status— if not logged in, ask them to run! gh auth login, then re-check.- Resolve the repo:
gh repo view --json nameWithOwner -q .nameWithOwner(run from the repo, or pass--repo OWNER/NAME). Confirm it's a Polis repo before triggering anything —gh api repos/{owner}/{repo}/contents/scripts/pipeline.sh >/dev/nullmust succeed; if it doesn't, stop (wrong repo).
State panel
Build it from gh (read-only):
- Issues:
gh issue list --state open --json number,title,labels --limit 30. - PRs:
gh pr list --state open --json number,title,labels,headRefName,isDraft,reviewDecision --limit 30.
For each issue/PR show: number, title, current agent:* + status labels, the computed stage,
and the suggested next action from the stage graph below. Group by "needs you" (has a
status label awaiting a human) vs "in flight".
Stage graph (what to suggest next)
| Current signal | Stage | Suggested next |
|---|---|---|
| New issue, no PR, reads like a big idea | idea | agent:arch (plan) or agent:spec (single unit) |
arch-review label / open arch/* PR |
arch drafted | agent:rearch (revise) or agent:decompose (make issues) |
| Generated issue, no spec PR yet | ready to spec | agent:spec |
spec-review label / draft agent/* PR with a spec |
spec drafted | agent:respec (revise) or agent:code (implement) |
Open agent/* PR, needs-human-review, tests passing |
code ready | review feedback → merge, or agent:fix |
PR with tests-failing |
tests red | agent:fix (no comment = make tests pass) |
PR with agent:cap-reached |
review loop capped | read verdicts → agent:fix, or re-run AI review with agent:code_review[:<profile>], or merge |
agent:failed |
run errored | gh run view <id> --log-failed to see the error, fix the cause, then re-apply the original trigger label |
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.
- 2mo ago First seen · 122 lines · 81 tokens per session scan A c1c5e8143abc
drive is a skill published in the GitHub repository permit0-ai/permit0 (192 stars, last pushed 4mo ago), licensed Apache-2.0. It adds 81 tokens to every session and 1,855 once invoked, about $0.0003 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 2026-08-30.
Other skills, from other repositories
interpret-results
Interprets AIR Blackbox scan results and maps findings to specific EU AI Act articles, recitals, and remediation steps. Use when the user has scan output and wants to understand what to fix, why it matters, or how to prioritize.
compliance-scan
Scans a Python AI project for EU AI Act compliance gaps using AIR Blackbox. Use when the user asks to check compliance, scan their code, audit their AI project, or mentions EU AI Act, Articles 9-15, or compliance checking.
auto-build-loop
Autonomous build driver after SDD spec approval: phase-scoped headless build sessions, driver-owned commits, cross-provider review gates, milestone pauses, and fail-closed escalation — under explicit opt-in autonomy profiles.
deep-code-review
Use when reviewing current branch for bugs before pushing or merging, when wanting a thorough multi-agent review of local changes, or when preparing work for human review.
gh
GitHub API access and project management automation for the hallucination-detector repo. Uses octokit with proxy-aware client for all GitHub operations — issues, PRs, labels, milestones, Projects V2. No gh CLI required.
ship-verification
Post-merge verification pass for Grantex. After any PR merges to main or ships to prod, run this pass without being asked — verify README, Mintlify docs, landing page, SDK READMEs (TS, Python, Go), integration docs, and SEO surfaces are consistent with what just shipped. Use this skill whenever a PR merges, a release…