メニュー
Show the Second Brain's recent advisories with their evidence and adjudicated verdicts, plus the ones the gate dropped and why. Use when the user asks what the Second Brain said, why it said it, or why it stayed silent.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
SOC 職業分類に基づく
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。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/shofer-dev/shofer --skill second-brain-whyコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SKILL.md を表示中
View or change the Live Memory plugin's configuration for this workspace — the provider profile the memory LLM uses, retention caps, the external-edit watch glob, the compaction interval, and the context-window budget/threshold. Use when the user asks to point Live Memory at a different model/provider, tune how much it remembers, or change its context budget.
Provides context about the Shofer evals system structure in this monorepo. Use when tasks mention "evals", "evaluation", "eval runs", "eval exercises", or working with the evals infrastructure. Helps distinguish between the evals execution system (packages/evals, apps/web-evals) and the public website evals display page (apps/web-shofer-code/src/app/evals).
Show what the Second Brain has observed, spent and said — observed volume per segment, pass latency, window fill, token use and cost, advisories generated / gated / delivered, and uptake per detector. Use when the user asks whether the Second Brain is watching, what it has cost, or whether its advice is being taken.
| name | second-brain-why |
| description | Show the Second Brain's recent advisories with their evidence and adjudicated verdicts, plus the ones the gate dropped and why. Use when the user asks what the Second Brain said, why it said it, or why it stayed silent. |
Call the plugin request why (plugin second-brain, params { taskId? }).
Relay, per task:
adopted | partially_adopted | rejected | already_handled | no_evidence | contradicted, or open). Outcomes are self-reported by the
detector from what it observed AFTER delivery, evidence-required, defaulting to
no_evidence — uptake, never impact.duplicate, rate_limited,
stale, below_floor, suppressed, muted, expiry). The gate's decisions
must be inspectable or nobody will trust the channel.