Determines if an interaction deserves persistence in the project state based on defined rules.

Install

mkdir -p .claude/skills/ops-judge && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/12848" && unzip -o skill.zip -d .claude/skills/ops-judge && rm skill.zip

Installs to .claude/skills/ops-judge

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

仅在主 agent 显式调用时使用,判断"刚完成的一轮有效交互"是否值得写入 `.codex/state/`。只负责给出 ignore/defer/apply 结论和最小事件负载,不负责写文件。
97 charsno explicit “when” trigger
Advanced

Key capabilities

  • Determine if an interaction should be ignored
  • Decide if an interaction should be deferred
  • Conclude if an interaction should be applied to state
  • Generate a minimal event payload for state updates
  • Identify changes in Focus, Next 3, Blockers, Active Issues, or Done
  • Recognize new or confirmed key decisions

How it works

The skill evaluates a recent interaction based on predefined rules to decide whether it should be ignored, deferred, or applied to the project state, then generates a minimal event payload.

Inputs & outputs

You give it
A round of effective interaction
You get back
A JSON object with a decision (ignore, defer, apply), reason, summary, and event payload

When to use ops-judge

  • Filter noise from interaction logs
  • Identify actionable outcomes
  • Determine if a change warrants state persistence
  • Sync state intelligently

About this skill

Ops Judge

你现在执行的是"状态记录判定"。

目标:只基于"刚刚这轮有效交互"判断是否应写入项目状态,不直接写任何文件。

这是主 agent 协议中的显式步骤,不是平台 lifecycle hook。

除非在单独调试判定逻辑,否则不要把你当作默认入口;正常流程统一先走 ops-sync

判定原则

优先判断"为什么不该记",再判断"为什么该记"。

三态规则表

判为 ignore

  • 纯讨论、寒暄、重复确认
  • 只读文件但没有形成新结论
  • 小命令成功执行,但没有改变当前状态判断
  • 临时试探失败,且没有沉淀出可复用经验
  • 仅修改排版、措辞、注释格式
  • 子 agent 返回了内容,但主 agent 明确认为"当前不采纳,且后续也不需要合并"

判为 defer

  • 有一点新信息,但尚不足以单独写入
  • 连续小步操作,适合在阶段结束时合并
  • 子 agent 有中间结论,但主 agent 还未正式采纳
  • 小步文件修改已发生,但还没有形成本轮可确认结论
  • 命令结果或外部检索给了线索,但尚未改变 Focus、Next 3、Blockers 或 Active Issues

判为 apply

  • 当前 Focus、Next 3、Blockers、Active Issues、Done 发生变化
  • 新增或确认了关键决策
  • 修改了文件且已形成明确结果
  • 命令输出改变了判断
  • 外部检索得到已采纳的新结论
  • 定位了根因,或排除了关键假设
  • 产生了可复用的轻量 KB

冲突时优先级

  • 若同时满足 applydefer,优先 apply
  • 若同时满足 deferignore,优先 defer

输出要求

优先只输出一个原始 JSON 对象。若客户端自动包裹代码块,后续 normalize 层会负责提取;你不要主动添加解释性正文。

字段固定如下:

{
  "decision": "ignore | defer | apply",
  "reason": "一句话原因",
  "summary": "本轮发生了什么",
  "event": {
    "source": "ops-judge",
    "intent": "本轮意图",
    "result": "本轮结果",
    "tags": ["tag1"],
    "links": ["ISSUE-001"],
    "focus": "可选",
    "blockers": ["可选"],
    "next3": ["可选"],
    "active_issues": ["可选"],
    "done_add": ["可选"],
    "notes": ["可选"],
    "issue_id": "可选",
    "issue_status": "open|doing|blocked|done",
    "issue_debug_add": ["可选"],
    "kb_id": "可选",
    "kb_pattern": "可选",
    "kb_root_cause": "可选",
    "kb_fix": ["可选"],
    "kb_precheck": ["可选"],
    "brief_append": "可选"
  }
}

补充要求:

  • ignore 时,event 保持最小化,只保留 intentresult
  • defer 时,不要伪造完整 state 更新;只保留后续合并需要的最小信息
  • apply 时,只填写本轮已确认的信息;缺失字段直接省略
  • deferapply 时,event.source 固定写 ops-judge
  • 不要输出解释性正文,不要写 Markdown 列表
  • 不要自由发明 dedupe_key、事件路径或状态文件路径

ops-apply 的关系

  • 你不负责写文件
  • ops-sync 会负责调用你、规范化你的输出、并决定是否唤醒 ops-apply

When not to use it

  • When the user needs to write files directly
  • When the user needs to debug judgment logic in isolation

Limitations

  • The skill does not write files directly
  • The skill's judgment is based on predefined three-state rules
  • The skill is intended to be called by the main agent

How it compares

This skill provides a structured, rule-based judgment for state updates, ensuring consistency and minimizing unnecessary writes, unlike ad-hoc decisions about what to record.

Compared to similar skills

ops-judge side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
ops-judge (this skill)03moNo flagsAdvanced
using-superpowers953moNo flagsBeginner
ultrawork112moNo flagsAdvanced
clawhub252moReviewIntermediate

Try saying

Example prompts that trigger this skill in your AI assistant.

You might also like

using-superpowers

obra

Use when starting any conversation - establishes mandatory workflows for finding and using skills, including using Skill tool before announcing usage, following brainstorming before coding, and creating TodoWrite todos for checklists

95205

ultrawork

Yeachan-Heo

Parallel execution engine for high-throughput task completion

11184

clawhub

openclaw

Use the ClawHub CLI to search, install, update, and publish agent skills from clawhub.com. Use when you need to fetch new skills on the fly, sync installed skills to latest or a specific version, or publish new/updated skill folders with the npm-installed clawhub CLI.

25151

skill-installer

openai

Install Codex skills into $CODEX_HOME/skills from a curated list or a GitHub repo path. Use when a user asks to list installable skills, install a curated skill, or install a skill from another repo (including private repos).

29141

continuous-learning

affaan-m

Automatically extract reusable patterns from Claude Code sessions and save them as learned skills for future use.

995

memory-keeper-proactive-context-maintenance

b4CU-R4U

Automatically detect and maintain memory freshness by monitoring context staleness, significant code changes, task completions, and phase transitions. Proactively suggests and executes memory sync operations with user confirmation. Use when the user says "sync memory", "update context", or when the Skill detects that context is stale (>2 hours), significant changes have occurred (new commits), tasks completed, or major milestones reached. Replaces passive "context is stale" warnings with active maintenance.

694

Search skills

Search the agent skills registry