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 agentmods add skills/winterddo/claude-code-skill-autopilot/tutornpx skills add WinterDDo/claude-code-skill-autopilot --skill tutorgit clone --depth 1 https://github.com/WinterDDo/claude-code-skill-autopilotWrote 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/winterddo/claude-code-skill-autopilot/tutor)<a href="https://agentmods.dev/skills/winterddo/claude-code-skill-autopilot/tutor"><img src="https://agentmods.dev/badge/skills/winterddo/claude-code-skill-autopilot/tutor.svg" alt="Measured on agentmods" 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 | $0.00091 | $0.00492 |
| Opus 5 | $0.00046 | $0.00246 |
| Sonnet 5 | $0.00018 | $0.00098 |
| Haiku 4.5 | $0.00009 | $0.00049 |
Grade B, and why
tutor scanned grade B with 1 finding 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 4d 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.
Strips warnings and disclaimersmediumAnti-refusal
Omitting safety caveats hides risk from the user and is a common jailbreak preamble.
Run a short interactive tour. Speak the user's language. Keep every step to 2-3 sentences. Never lecture. What it actually says
Skill Autopilot tour (2 minutes, hands-on)
Run a short interactive tour. Speak the user's language. Keep every step to 2-3 sentences. Never lecture.
Step 1 — the magic moment (auto plan mode)
Ask the user to type a deliberately big request, for example: "design and build a statistics feature for this project" (or let them invent one). When they send it, Claude enters plan mode by itself before touching any file. Tell them to REJECT the plan afterwards: nothing changed, and they just watched the autopilot think before acting.
Step 2 — the undo button
Create a throwaway file (test-autopilot.md, three lines). Then tell the user: "say 'undo that'". The autopilot's first reaction is to hand them /rewind (press Esc twice) instead of patching forward. Let them open the rewind menu and look — selecting nothing changes nothing.
Step 3 — the habit card
Render this table, translated to the user's language:
| Command | When | How |
|---|---|---|
| /clear | Starting an unrelated new task | type /clear — new task carries no old baggage |
| /btw | A side question mid-task | type /btw <question> — zero pollution of the main thread |
| /rewind | Dissatisfied with a change | press Esc twice — back to before the damage |
| plan mode | Any big change | press Shift+Tab — review the plan before any file changes |
Close with one line: "Everything else you don't need to learn. What Claude can do, it just does; what only you can press, it hands you before the moment — with the reason why."
Step 4 — wrap
Mention once: the profile skill shows what the autopilot did for them; the config skill mutes or quiets it in one word.
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.
- 4d ago First seen · 30 lines · 91 tokens per session scan B 3165047699d1
tutor is a skill published in the GitHub repository WinterDDo/claude-code-skill-autopilot (6 stars, last pushed 2mo ago), licensed MIT. It adds 91 tokens to every session and 492 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it B with 1 finding (strips warnings and disclaimers). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
platform-skills
Use when troubleshooting, implementing, reviewing, or auditing platform infrastructure as a system — where Kubernetes, GitOps, CI/CD, and security concerns intersect. Provides structured diagnosis with blast radius, validation steps, and rollback plan for: Kubernetes, Flux CD, Argo CD, Terraform, GitHub Actions…
notebook-paired-with-prose
Use when drafting or executing a paired notebook for a bookwright chapter. Covers when notebooks are required vs optional, numerical-sanity-target convention, exec-from-fresh-kernel requirement, and stack-specific execution commands (Jupyter/R Markdown/Quarto).
code-explainer
Explains codebases visually with ASCII architecture diagrams, code flow tracing, and interactive deep-dives. TRIGGER THIS SKILL when users want to understand, learn about, or get oriented in a codebase — whether they say "explain this codebase", "what does this project do", "how does this work", "walk me through the…
guided-learning
Adaptive guided learning skill that teaches any concept by tracing backward to prerequisites, assessing the learner's level, and dynamically switching between Socratic, Bloom's taxonomy, and Visual teaching styles based on responses. Uses web search for accuracy and research papers via alphaxiv when relevant. Produces…
skill-audit
Audit codebases for quality, consistency, and broken patterns — use for pre-release or tech debt review.
skill-context-detection
Auto-detect work context (Dev vs Knowledge) — use to tailor workflows based on current task type.