audit-agency-continuity
Run a read-only continuity and evidence audit of a persistent or autonomous agent workspace. Use when Codex must assess whether goals, memory, corre…
它会碰到什么
扫了多少3 个文本文件,11 KB
它会碰到什么改身份文件读文件
命中总数2 处
命中统计严重 0 · 高 1 · 中 1 · 低 0
逐条看命中(1 条严重或高危)
- 高
scripts/audit_continuity.py:14identity-writeOBJECTIVE_FILES = ("MANDATE.md", "CHARACTER.md", "AGENTS.md")
这一栏是扫描器报的事实,不是结论。命中多不等于有毒(安全工具、规则库、示例脚本本来就会包含危险写法),命中少也不等于干净。它和你手上的凭据、文件、网络有什么关系,需要你自己看。
技能内容
Audit Agency Continuity
Run the deterministic auditor before interpreting narrative documents:
python3 scripts/audit_continuity.py --project <workspace> --state-dir <state-directory>
Use --json for machine consumption. The command is read-only.
Then:
- Treat
failas a contradicted or missing durability requirement. - Treat
warnas unproven, not achieved. - Inspect cited paths only for consequential findings; do not load an entire Markdown corpus by default.
- Separate operational continuity from commercial success. A healthy scheduler or memory database never proves revenue.
- Never repair findings unless the user asked for changes. For a requested repair, preserve existing state, add tests, and rerun the audit.
- Report the smallest set of decisive findings: what survives restart, what can silently drift, what requires human authority, and which claimed outcome lacks external evidence.
Do not infer that numerous documents equal memory, that a timer equals successful execution, that model agreement equals external evidence, or that gross or pending revenue proves owner-available contribution.
想直接用这个技能?
本站把开放许可(MIT / Apache 等)的技能按仓库打包整理到网盘,点一下转存到你自己的网盘,不用一个个从 GitHub 拉。许可未声明的技能只给原始仓库链接,不打包。
它属于哪个仓库
星标★ 1,027
本站分层T1
该仓技能数1910
原文件路径
plugins/revertcreations/agency-continuity-audit/skills/audit-agency-continuity/SKILL.md