一键导入
deputy-briefing-format
How a deputy should structure its briefing XML — what to include, what to leave out, how to weight relevance.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
How a deputy should structure its briefing XML — what to include, what to leave out, how to weight relevance.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
How to orchestrate work across large documents and multi-agent delegations without burning token budgets. Reconnaissance-first, search-then-read, parallel-after-reconnaissance.
Decision criteria for delegating to the writer agent vs synthesising directly in your own response.
When and how to use the document library
How the Coder agent drives an OpenCode session via dispatch_coder — the engineering loop, when to dispatch, and how to verify the result is real (vs. claimed).
Maintain a personal index of interesting things, open questions, and follow-ups across sessions
When and how to render a promoted export to PDF via the export_pdf tool. Reads the writer's draft → export → PDF lifecycle and the operator-facing failure modes when the host is missing pandoc or tectonic.
| name | deputy-briefing-format |
| description | How a deputy should structure its briefing XML — what to include, what to leave out, how to weight relevance. |
| agents | deputy |
Your output is a <deputy_briefing> XML block. Its structure:
<deputy_briefing>
<signal>high_novelty</signal>
<relevant_cases>
<case>
<case_id>CBR-042</case_id>
<similarity>0.87</similarity>
<summary>Short description of how the prior case applied and what the lesson was.</summary>
</case>
</relevant_cases>
<relevant_facts>
<fact>
<key>cycle_log_test_pattern</key>
<value>One-line answer to what this fact says.</value>
</fact>
</relevant_facts>
<known_pitfalls>Plain-prose note about a recent pattern of failure that the agent should be aware of.</known_pitfalls>
</deputy_briefing>
known_pitfalls is optional. Only populate it if recent narrative
entries show a clear pattern worth warning about.An empty briefing (signal=silent, no cases, no facts) is a correct and
valuable output when nothing in memory applies. Don't pad.
Your similarity scores are your honest estimate of how applicable a prior case is to the current delegation:
If you cite a case below 0.5, explain in the summary why it's still useful.
routine — known pattern, similar delegations have succeeded; the agent
doesn't need special guidancehigh_novelty — this looks unfamiliar; cog should be awareanomaly — there's something off (repeated failures in this domain,
conflicting evidence)silent — nothing in memory applies; use this freely, it's not a failureEach case summary should be one sentence, specific to the lesson:
Good:
Bad:
& → &, < → <, etc.Return:
<deputy_briefing>
<signal>silent</signal>
</deputy_briefing>
The specialist will proceed without briefing context. This is not a failure; it is the honest answer when memory doesn't cover the situation.