Agent Status

Agent status

Target agent: $ARGUMENTS

Step 1 — Validate

• If no argument given → stop and ask: "Which agent? Run /list-agents to see available agents."
• Glob .claude/agents// — if directory doesn't exist, stop and say agent not found.
• This skill only works for Type B autonomous agents (folder structure). Type A subagents have no journal or outputs.

Step 2 — Read agent state

Read the following files in order:

1. .claude/agents//AGENT.md — extract: Mission, Goals & KPIs table
2. .claude/agents//MEMORY.md — extract: "What Works" and "Last Updated" sections
3. .claude/agents//outputs/ — Glob for all files, sort by name (date-prefixed), take the 3 most recent
4. journal/ — Glob for all files, sort by name, take the most recent entry that mentions this agent name

Step 3 — Summarize

Output:

===  status ===

Mission: 

KPI Progress:
  •  | KPI:  | Baseline:  | Target:  | Current: 

Last activity:
  • Journal:  — 
  • Last output:  ()

Memory (what's working):
  

Last updated: 

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."

Notes

• Do NOT read all journal files — only the most recent one that mentions the agent name
• Current KPI value comes from the most recent output file or journal entry — if not found, show "unknown"

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: enbiyagoral
• Source: enbiyagoral/.claude
• License: MIT

Install and usage instructions live in the source repository linked above.