code-worker-governance — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited code-worker-governance (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
code-worker MCP,或当前任务明显适合交给 code-worker 时启用。1. 快速2. 思考3. 自动1. 快速 -> execution_tier=fast2. 思考 -> execution_tier=think3. 自动 -> execution_tier=auto快速/思考/自动,不默认展开底层模型名。fastthinkauto 只允许存在于“用户选择”和“主线程编排”阶段,不作为多子任务最终汇总的展示值。快速(flash)思考(pro)快速/思考3. 自动:快速:思考:纯读大范围 且决定起 worker:思考思考纯读小范围纯读大范围小范围改单文件多文件实现调试闭环长时间后台任务纯读小范围 或 小范围改单文件:纯读大范围:review 倾向的 task,接受 completed_no_changes多文件实现 或 调试闭环:长时间后台任务:code_worker_implement_in_workspacecode_worker_start_implementationget_job / wait_for_jobinclude_logs: trueinclude_events: trueinclude_diff: truecode_worker_wait_for_job 只用于短前台观察,不作为主循环。max_wait_ms=180000~240000status=runningprocess_alive=truelast_error_kind 为空token_progressing=true 或 stream_progressing=truewait_for_jobnext_recommended_check_after_ms 调 code_worker_get_job300000~600000ms 读一次context-compression 做摘要收口liveness_state=running_suspected_stallliveness_state=running_stalledlast_error_kind 非空token_progressing:text_delta / final_result 附近:fast_patchscaffold_or_testsdebug_loopagentic_codingcomplex_reasoning下发 worker 时,主线程只给最小必要上下文,建议结构:
Task:
- Task label: <短标签,供最终汇总>
- Goal: <单一明确目标>
- Requested tier: <用户选择 fast/think/auto>
- Resolved tier: <本 worker 实际使用 fast/think>
- Scope: <允许改动的目录/文件>
- Do not touch: <禁止路径>
- Validate: <检查命令>
- Output: <希望 worker 最终说明什么>
- Previous result: <如有续跑,仅带 job_id / 终态 / 当前 diff 摘要>Summary
- Goal completed: <yes/no>
- Files changed: <list>
- Checks run: <list + result>
- Residual debt: <only if any>
- Risk points: <only if any>
- Contract-impacting changes: <only if any>log.error如果当前仓库有明确研发规范,task brief 里追加一段:
Implementation constraints:
- <规则 1>
- <规则 2>
- <规则 3>例如:
Implementation constraints:
- 命名统一业务术语,不混用 renew/extend
- 核心步骤前写简短注释
- 失败统一 log.error,带关键标识
- 不做无关重构L1 轻审,只有命中高风险才升级 L2 重审。statustask_labelexecution_tier / model_labelfailure_reasonfiles_changedreview_summarychecks_runstdout_tailstderr_tailrecent_eventstool_eventsL1 轻审 默认只看:files_changedreview_summarychecks_run~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.