Use when manage multiple local CLI agents via tmux sessions (start/stop/monitor/assign) with cron-friendly scheduling.
日本語の概要は準備中です。原文の説明を表示しています。
Use when the user wants to share or back up learned instincts — validates the local instincts file, allowlists the six known fields, strips secrets and home-directory paths, and writes a sanitized JSON artifact the user names. Never modifies the instincts file itself.
インストールする前に、エージェントに与えられる指示の中身を確認できます。
Concept inspired by the instinct/skill-create ideas in affaan-m/ECC; no code copied.
Instincts are small learned behaviours stored locally under
~/.aos/instincts/instincts.json (format defined in /instinct-status).
Exporting one means making it portable — into version control, a teammate's
machine, a dotfile repo. That is exactly where leaks happen, so sanitization
is the job, not a nicety.
Read ~/.aos/instincts/instincts.json. If it is missing, the file is not
silently treated as empty: report that there is nothing to export and stop.
If the JSON does not parse, fail loudly with the parse error — never export
garbage "with best effort".
Shared bounds, same as on the import side: trigger and action must be
JSON strings ≤ 200 / ≤ 400 characters, no newlines or control characters;
at most 200 instinct entries. A store that fails these bounds is corrupt, not
exportable — fail loudly with the reason the same way as the parse-error path
above, never ship an artifact /instinct-import would reject wholesale.
Field allowlist: every exported entry keeps exactly the six shared fields
— id, trigger, action, confidence (0–1), source, updated — and
every other field is dropped. No auxiliary metadata travels with the
entry. Nothing on the allowlist is negotiable; extra fields are removed, not
published.
For every entry, scrub trigger and action against these redaction rules
(a hit means remove the content, or the whole entry if the lesson dies with
it):
sk-… prefixed keys.ghp_… tokens.github_pat_… tokens.xox-prefixed Slack tokens (xoxp-, xoxb-, …).AKIA… access-key ids.header.payload.signature).user:pass@… in any URL./Users/<name>/…, /home/<name>/…,
C:\Users\<name>\…. Rewrite to ~/…
or describe the location in words. Relative paths (src/lib/) are safe.An instinct that only exists to memorialize a secret is dropped, not partially redacted. A missing entry is recoverable; a leaked credential is not.
After redaction and field-filtering, serialize the JSON and scan the whole serialized output once more against the section-2 list before any write. Redaction that happens only on the entry objects can miss paths, e-mails or tokens introduced by formatting or by fields added late; the final scan is what closes that gap. If the re-scan still finds a hit, fix it and re-scan again — never write with a known hit remaining.
Default if the user names no path: a scratch/output folder the user names —
for example ~/instinct-export/instincts-YYYY-MM-DD.json. Never the home
root (~/instincts-export-….json is out), never a dotfiles repo path, never
inside ~/.aos/. If the user names a path inside the AOS state dir, refuse
it — resolve .. segments and symlinks first and check the resolved path,
not the spelled-out one. Never overwrite an existing export file without the
user's explicit yes. The export keeps version: 1 and an exported: <ISO date> key at the top level; the instincts file itself is never
modified.
Summarize: N instincts exported, K sanitized (list the ids), M omitted and
why (one line each). Show the output path and state that the original file
was untouched.
まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Use when manage multiple local CLI agents via tmux sessions (start/stop/monitor/assign) with cron-friendly scheduling.
日本語の概要は準備中です。原文の説明を表示しています。
Use when a hybrid memory system that provides persistent, searchable knowledge management for AI agents (Architecture, Patterns, Decisions).
日本語の概要は準備中です。原文の説明を表示しています。
Reference for how BDB structures autonomous software engineering work — the seven-node dispatcher graph (Architect, TechLead, UI/UX, Engineering, Media/EventTech, Reviewer, Shipping) that /startcycle-graph actually runs. Use when you need the high-level lifecycle framing without inventing your own process.
日本語の概要は準備中です。原文の説明を表示しています。
Tools are how AI agents interact with the world. A well-designed tool is the difference between an agent that works and one that hallucinates, fails silently, or costs 10x more tokens than necessary. This skill covers tool design from schema to error handling.
日本語の概要は準備中です。原文の説明を表示しています。
Harness patterns for coding agents — memory, permissions, context engineering, delegation, skills, hooks, bootstrap.
日本語の概要は準備中です。原文の説明を表示しています。
agenttrail: live map of a multi-agent build in the browser: which plan component is being worked on, by which agent or harness, what is done and what is stuck. Use when a multi-agent pipeline starts (/startcycle, /startcycle-graph, /teamwork-preview) or after a plan-canvas approve, when the user asks to see what the agents are doing ("live map", "build board", "who is running now"), or by yourself whenever a multi-agent run is under way — no user prompt needed.
日本語の概要は準備中です。原文の説明を表示しています。