メニュー
Defines how to use wiki tools (wiki_get, wiki_write) and — critically — how to surface wiki usage to the user. Load whenever you read, write, or might benefit from an article.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
How to use Aperio's memory tools well — when to recall, what is worth remembering, how to correct a wrong memory, and what must never be stored. Read this when connected to an Aperio MCP server; the tool schemas describe parameters, this describes judgment.
Use this skill to iteratively tune the keywords and descriptions of other SKILL.md files so the skill matcher fires the right skill for a prompt. Runs an autonomous mutate → score → keep/discard loop against a frozen eval set (exam §7 skill triggering), advancing a git branch when match accuracy improves. The eval and scorer are read-only ground truth; the curated keyword frontmatter of other skills is the only editable surface.
Use this skill before writing new code — when the ask is to add a helper, a small feature, a utility, or a new module, and the right amount of code has not been decided yet. Triggers: "write me a helper for…", "add a small feature", "do I need a library for this", "keep it minimal", weighing a new dependency, or any request that could be met by something that already exists. Walk the ladder first: reuse beats stdlib beats platform beats dependency beats new code. This is the pre-write sibling of code-simplification, which cleans up afterwards.
Use this skill when reading from or writing to the persistent memory store. Covers the memory tool API (recall, remember, update_memory, forget) and — for advanced direct access — the SQL schema, vector search patterns, and retrieval priority order against the Aperio Postgres store. Triggers: user states a preference, shares a fact or decision, gives a correction, asks the agent to remember something, or starts a task where past context would affect the output.
Use this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting/decrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill.
Toolkit for styling artifacts with a theme, and for creating distinctive, production-grade frontend interfaces. Use for: applying pre-set or custom themes to slides, docs, reports, HTML pages; building web components, pages, dashboards, React components, HTML/CSS layouts; or styling/beautifying any web UI with high design quality. ALWAYS use this skill before building any frontend or UI — it prevents the default "clean SaaS with purple gradient" trap that makes every AI-built site look identical.
| name | wiki |
| description | Defines how to use wiki tools (wiki_get, wiki_write) and — critically — how to surface wiki usage to the user. Load whenever you read, write, or might benefit from an article. |
The wiki is a cache of LLM-authored, cited synthesis articles over the user's memories.
Use it for composite or recurring topics (architecture overviews, project state, "what
do I know about X" questions). Do not use it for single-fact lookups — call recall directly.
Hybrid FTS + semantic search over existing articles. Always call this before wiki_write — if a
relevant article already exists, update it (same slug) instead of creating a duplicate. Returns slugs,
titles, summaries, and scores; no bodies. Stale matches are included but down-weighted.
Fetch an existing article. First line of the result is a breadcrumb:
🔖 From wiki: [[slug]] (rev N · status · updated YYYY-MM-DD)
If the article is stale and the user is asking for current truth, prefer
wiki_get(slug, refresh=true) over a manual recall+wiki_write loop — the server
will rewrite the article in-place using the configured refresh provider
(see WIKI_REFRESH_PROVIDER), which is typically a cheap/local model. The
returned body is the refreshed version. If refresh isn't configured the call
silently degrades to returning the stale body with a footer note.
Create or update an article. Workflow:
wiki_search(topic) — check whether an article already exists; if yes, reuse its slug.recall(topic) — gather candidate memories.body_md; cite memories inline as [[mem:<uuid>]] and link siblings as [[other-slug]].wiki_write with the cited memory ids in source_memory_ids.When you used a wiki article — fully or partially — to answer the user, copy the breadcrumb line verbatim as the first line of your user-facing reply, before any other prose.
Example:
🔖 From wiki: [[aperio-architecture]] (rev 3 · fresh · updated 2026-05-17)
Aperio runs as a Docker stack with Postgres + pgvector behind an MCP server…
If you consulted multiple articles, list each breadcrumb on its own line at the top.
If the article was stale and you regenerated it via wiki_write before answering, surface
the breadcrumb for the new revision (the one you just wrote), not the stale read.
Why: the user has no UI for the wiki yet. The breadcrumb is the only signal that the feature exists and is working. Dropping it silently makes the wiki invisible.
Write proactively when you notice you've stitched together ≥3 memories on the same topic to answer a question, and the topic is likely to come up again. Confirm slug with the user on first creation; slugs are immutable once other articles link to them.
body_md that aren't grounded in a source_memory_ids row.wiki_get returns status: stale and the user wants current truth, regenerate before
answering.