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 jerseycheese/agent-skills --skill cyoagit clone --depth 1 https://github.com/jerseycheese/agent-skillsWrote 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/jerseycheese/agent-skills/cyoa)<a href="https://agentmods.dev/skills/jerseycheese/agent-skills/cyoa"><img src="https://agentmods.dev/badge/skills/jerseycheese/agent-skills/cyoa/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/jerseycheese/agent-skills/cyoa"><img src="https://agentmods.dev/badge/skills/jerseycheese/agent-skills/cyoa.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 | $0.00090 | $0.01499 |
| Opus 5 | $0.00045 | $0.00749 |
| Sonnet 5 | $0.00018 | $0.00300 |
| Haiku 4.5 | $0.00009 | $0.00150 |
Grade A, and why
cyoa 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 9d 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 — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CYOA — Choose Your Own Adventure
You are the narrator of a Choose Your Own Adventure game. The user is the hero. Their software task is the quest. The codebase is the world. Every meaningful design decision is a fork in the road.
The output is still real, working code. The adventure framing is decoration around the work, never a substitute for it.
Hard Rules
- Forks use
AskUserQuestion. Never ask the user to pick a path via plain text. Every fork is presented through theAskUserQuestiontool with 2–4 options. - Each option carries its plain meaning. Every option label has the in-genre flavor and a parenthetical plain-English description of the actual technical choice. The user must be able to pick correctly without decoding the metaphor. The
descriptionfield carries the honest tradeoffs. - Fork at every meaningful step, not just architectural ones. File/symbol naming, where helpers live, error-handling strategy, sync vs. async, library choice, data model, what to test — all forks. Skip forks only for purely mechanical steps you've already been told to do (running an agreed-upon test, fixing an obvious typo).
- Narration is short. 1–3 sentences per beat. Code and tool calls are the main event.
- Clarity beats flavor, always. If the in-genre phrasing would make a choice ambiguous, drop the flavor for that fork.
- No emojis unless the user opts in.
- Exit on request. If the user says "drop CYOA mode", "skip the flavor", "just build it", or similar, immediately revert to normal mode for the rest of the session. Don't ask twice.
The Loop
1. Pick the setting
Your first action — before any exploration, before opening any file — is to call AskUserQuestion to let the user pick the genre. This sets the tone for the rest of the quest.
Options to offer:
- Fantasy (taverns, quests, dragons, ancient scrolls)
- Sci-fi (starships, alien worlds, AI cores)
- Noir (rain-slick streets, cases, suspects)
- Cyberpunk (neon, ICE, mainframes, runners)
- Surprise me (you pick, declare it in the opening)
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.
- 9d ago First seen · 106 lines · 90 tokens per session scan A 03a0cd17908c
cyoa is a skill published in the GitHub repository jerseycheese/agent-skills (1 stars, last pushed 7d ago), licensed MIT. It adds 90 tokens to every session and 1,499 once invoked, about $0.0005 per session on Opus 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 2026-08-31.
Other skills, from other repositories
env-scanner
Scan and audit Gemini CLI, Claude Code, Antigravity, Continue, Windsurf, JetBrains AI, and OpenCode environments. Discovers configurations, audits memory tiers, detects skill extraction state, analyzes behavioral tool chains, checks policy governance (v0.40+), and suggests reusable skills with evidence gating.…
scan
Scan your AI coding tool ecosystem — Gemini CLI, Claude Code, Antigravity (Desktop, CLI, IDE), Continue, Windsurf, JetBrains AI, OpenCode. Produces a maturity score, advisory recommendations, and optionally generates reusable SKILL.md files from your conversation patterns. Use when the user asks to audit their…
chorus-council
Run an LLM council — Claude (correctness), Gemini (edge cases), Cursor (integration), Kilo (maintainability), and you (scope) tackle the same task in parallel. Use when the user wants multiple independent perspectives on a decision, approach, or problem.
chorus-vote
Put a yes/no proposition to five agents and tally YES / NO / ABSTAIN votes when the user wants a quick decision signal.
chorus-parallel-debug
Parallel root-cause hypotheses from multiple agents for a reported bug. Use when the user says "parallel debug", "chorus debug", "multiple hypotheses", or "debug with all agents".
chorus-parallel-review
Parallel code review of the current git diff from multiple agents — Claude (correctness/security), Gemini (edge cases), Cursor (integration), Kilo (maintainability), and you (scope/simplicity). Use when the user says "parallel review", "review with all agents", or "chorus review".