菜单
Show the current status of a specific autonomous agent — last journal entry, KPI progress, and recent outputs. Use when the user says "agent status", "how is <agent> doing", "<agent> durumu", or "agent-status <name>".
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
Analyze accumulated feedback memories and turn recurring patterns into permanent controls (rules, permissions, hooks) or common-mistakes entries. Use when the user says "improve", "iyileştir", "apply feedback", "kurallaştır", or "turn feedback into rules".
Project onboarding and template setup. Use when the user says onboard, setup, initialize, init, configure this project, or first-time setup. Guides through detecting project mode and stack, configuring CLAUDE.md, permissions, hooks, rules, and architecture docs.
Save session learnings, patterns, and corrections to persistent memory. Use when the user says "save learnings", "save memory", "remember this session", or at the end of a productive session.
Add a new Claude Code control following project conventions. Use when the user says "new rule", "add rule", "create rule", "yeni rule ekle", or wants to enforce behavior consistently. Guides through choosing enforcement surfaces, writing effective rule files, and keeping controls consistent.
Create a new Claude Code skill following project conventions. Use when the user says "new skill", "add skill", "create skill", or "yeni skill ekle". Guides through gathering requirements, writing effective SKILL.md files, and structuring skill directories.
Add a new Claude Code hook — shell script wired into settings.json. Use when the user says "new hook", "add hook", "yeni hook ekle", or wants to automate something that runs on every tool use, session start, or stop.
| name | agent-status |
| description | Show the current status of a specific autonomous agent — last journal entry, KPI progress, and recent outputs. Use when the user says "agent status", "how is <agent> doing", "<agent> durumu", or "agent-status <name>". |
| argument-hint | <agent-name> |
| allowed-tools | Read, Glob |
| disable-model-invocation | true |
Target agent: $ARGUMENTS
/list-agents to see available agents.".claude/agents/<agent-name>/ — if directory doesn't exist, stop and say agent not found.Read the following files in order:
.claude/agents/<agent-name>/AGENT.md — extract: Mission, Goals & KPIs table.claude/agents/<agent-name>/MEMORY.md — extract: "What Works" and "Last Updated" sections.claude/agents/<agent-name>/outputs/ — Glob for all files, sort by name (date-prefixed), take the 3 most recentjournal/ — Glob for all files, sort by name, take the most recent entry that mentions this agent nameOutput:
=== <agent-name> status ===
Mission: <mission>
KPI Progress:
• <Goal> | KPI: <metric> | Baseline: <X> | Target: <Y> | Current: <Z or "unknown — no data yet">
Last activity:
• Journal: <date> — <one-line summary of what happened>
• Last output: <filename> (<date>)
Memory (what's working):
<What Works section content, or "No patterns recorded yet">
Last updated: <date from MEMORY.md>
If journal/ is empty or has no entries mentioning this agent → print: "No journal entries found. Has the first cycle run yet?"
If outputs/ is empty → print: "No outputs generated yet."