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/learn-deck/learndeck/learn-patchquestnpx skills add learn-deck/learndeck --skill learn-patchquestgit clone --depth 1 https://github.com/learn-deck/learndeckWhat 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 | $0.00036 | $0.00675 |
| Opus 5 | $0.00018 | $0.00338 |
| Sonnet 5 | $0.00007 | $0.00135 |
| Haiku 4.5 | $0.00004 | $0.00068 |
Grade A, and why
learn-patchquest 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 yesterday.
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 — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Learn PatchQuest
PatchQuest has one shared local learning record: its browser UI and stdio MCP server use the same SQLite database. The learner selects a seeded course and path, then submits answers in the UI; use MCP to guide and evaluate those visible submissions.
Start or resume
- Read
AGENTS.md,README.md, anddocs/mcp.md. - If the UI is not running, ask the learner to run
bun run appand open the displayed local URL. Do not start it silently. - Ask the learner to use the UI's explicit Connect action for a supported host,
or continue without an agent. Then use
patchquest_list_courses; if the learner said only “Let's start,” ask them to choose a visible course, path, and workspace/context in the UI. - Use
patchquest_list_pathswith that course ID. If the learner has no path, let them create it in the UI or callpatchquest_create_pathonly after they explicitly choose the course, path, and workspace/context. - For a coding course, run only the source-documented read-only dependency checks. Report present and missing requirements; never install dependencies. When the checks pass, suggest the matching development command for the learner to run themselves.
- Use
patchquest_get_next_activityfor exactly one current section and question. Do not present a module menu.
One learning turn
- Read only the named source and the current section needed for this question.
- Ask the learner to inspect, explain, or make one bounded change inside their selected workspace. They own code and command execution.
- Record learner-reported code paths or command results through
patchquest_record_evidence. - Ask them to submit the displayed question in the browser with confidence.
- Read
patchquest_get_progress, locate the newsubmittedattempt, and evaluate it throughpatchquest_evaluate_answer. - Tell the learner that the feedback and status are visible in the UI. For a partial or incorrect answer, identify the exact gap, point to the source, and ask for a separate revision submission; never overwrite an old attempt.
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.
- yesterday First seen · 66 lines · 36 tokens per session scan A 5902b1393a6e
learn-patchquest is a skill published in the GitHub repository learn-deck/learndeck (0 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 36 tokens to every session and 675 once invoked, about $0.0002 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
autobrowse
Learn a reliable browser workflow by iterating on a real web task, recording strategy, and proposing a reusable skill.
yc-cli
Operator's guide to the YC CLI (yc) for the Gini batch demo — scoped to the yc commands the demo actually uses: the validated browser-forward login flow (tmux + browserconnect) and investor research against Bookface. Assumes yc is installed but NOT logged in yet. Load before staging or running the demo.
wrench
Use Wrench to give a bring-your-own coding agent bounded, local-first web capabilities: capture or scrape public and signed-in web pages into Markdown; archive authorized audio, video, and transcripts with verification; query encrypted cached email, contacts, inbox, and messaging views; resolve a live conversation…
k-teach
Create or revise source-grounded lessons, exercises, learning pages, WeChat articles/drafts, and HTML slide decks. Use for teaching content, learning assets, official-account content, or PPT-style presentations.
browse-and-verify
Browse a URL with Playwright to validate links, find page sections, and extract content for quiz sources. Trigger: validate link, check this URL, find the section, browse page, what does this page say about, verify source.
planpage
Render a skill's plan, review gate, quiz, poll, or report as a beautiful, self-contained INTERACTIVE HTML page via the open-source planpage package (Preact → static HTML + local post-back). The user edits and annotates the plan in the browser, attaches screenshots, stages everything in a fixed sidebar, then hits Send…