菜单
Manage context as a finite resource. Compress between phases, reinforce objectives in long sessions, decay old detail. Use when sessions are long, multi-phase, or approaching limits.
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
Dispatch Codex to review all unreviewed repository work, then verify its correlated findings ticket or clean result. Use when requesting an independent Codex review, auditing all changes, checking work since the last review, or waiting for Codex findings. Trigger: dispatch codex review, independent review, review all changes, check codex findings, run review agent.
Always-on: when a project has tickets, work the frontier (first unblocked open ticket), mark done, update plan status.
Incremental and full-history repository review with resumable adoption batches. Use when reviewing new work, initializing review tracking, reviewing all work to date, resuming historical review, inspecting unreviewed commits, or reviewing changed tickets. Trigger: review new work, adopt repository, review all history, resume review, review marker, unreviewed commits.
End-of-session handoff that captures current state for the next session. Use when ending a work session, switching context, or before a long break.
Consolidate project artifacts — promote scratch to memory, deduplicate memory, process decisions, organize scripts, update steering/skills accuracy. Use periodically or when the project feels cluttered.
Run and interpret crew-research evals — run.sh flags, scores.jsonl fields, activation TPR/FPR verdicts, resuming interrupted runs. Use when running evals, reading eval results, diagnosing a failing definition, or resuming a dead run. Trigger: run the evals, eval results, scores.jsonl, activation test, TPR, delta threshold, resume the run, skip-completed, known gap.
| name | context-budget-awareness |
| description | Manage context as a finite resource. Compress between phases, reinforce objectives in long sessions, decay old detail. Use when sessions are long, multi-phase, or approaching limits. |
| metadata | {"type":"steering","invocation":"passive","practice":null} |
Context is finite. Compress at phase boundaries rather than carrying everything forward; incorporate sub-agent findings as compressed artifacts, not raw transcripts. (If the human offers to restart the session at a phase boundary, that's a good point to accept — but never stop or suggest a new session on your own account.)
When compressing prior work, preserve decisions and drop mechanics:
| Age | Keep | Drop |
|---|---|---|
| Current phase | Full detail (files, diffs, decisions, blockers) | Nothing |
| Previous phase | Key outcomes + file paths + decisions | Exploration, intermediate reasoning, raw tool output |
| 2+ phases ago | One-line summary of what was decided | Everything else |
Never compress: decisions, constraints, acceptance criteria, file paths that are still relevant.
In long sessions (20+ messages), the model loses track of original objectives.
