agent-fitness — Claude Code skill: reports and installation
The agent-fitness Claude Code skill tracks sub-agent dispatch results and highlights blocked or failed runs. It responds to requests about agent calibration and performance trends.
MIT ★ 26 updated 2026-11-16
npx skills add wrg32786/aigent-os --skill agent-fitness -a claude-codeWhat agent-fitness does
The skill extracts dispatch records from JSONL and appends new rows to memory/AGENT_FITNESS.md. Run /agent-fitness to get dispatch totals, per-agent outcome counts, and recent trends. Add --days N to limit the report to a time window.
The report separates clean, blocked, errored, and partial outcomes. Calibration ratios exclude infrastructure errors, while repeat-blocker alerts appear after two or more blocks in the selected window. Classification relies on keyword matching; the skill does not reroute dispatches. Extraction runs through python3.
What it helps with
- compare clean and blocked dispatch counts across agents
- spot agents with repeated blocks in a chosen period
- find agent-and-task pairs that often end in failure
- check recent outcome trends for a sub-agent
Example prompts for the agent
- Extract new sub-agent dispatch records with
/agent-fitness extract. - Show the overall agent report with
/agent-fitness. - Limit the report to the last 7 days with
/agent-fitness --days 7.
How to install agent-fitness
1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill agent-fitness -a claude-code2Manually: put the files in place
~/.claude/skills/agent-fitness/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/agent-fitness/
1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill agent-fitness -a codex2Manually: put the files in place
.agents/skills/agent-fitness/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill agent-fitness -a cursor2Manually: put the files in place
.cursor/skills/agent-fitness/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install wrg32786/aigent-os skills/agent-fitness2Manually: put the files in place
.github/skills/agent-fitness/SKILL.md1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill agent-fitness -a gemini-cli2Platform CLI
gemini skills install https://github.com/wrg32786/aigent-os --path skills/agent-fitnessManually: put the files in place
.devin/skills/agent-fitness/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill agent-fitness -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about agent-fitness
How do I install agent-fitness in Claude Code?
Run npx skills add wrg32786/aigent-os --skill agent-fitness -a claude-code. The skill file is at ~/.claude/skills/agent-fitness/SKILL.md.
What does agent-fitness do?
It tracks sub-agent dispatch outcomes and produces per-agent reports. The results include blocks, errors, and recent trends.
What does agent-fitness need to run?
Extraction uses python3. Dispatch records come from JSONL, and the skill appends them to memory/AGENT_FITNESS.md.
Source text
SKILL.md · MIT
---
name: agent-fitness
agent: none
description: "Track sub-agent dispatch calibration from JSONL or ledgers. Use when checking agent fitness, clean/blocked/error trends, or which agents work."
allowed-tools: Bash, Read
user-invocable: true
triggers:
- agent fitness
- agent calibration
- agent stats
- sub-agent stats
- which agents work
- /agent-fitness
- agent dispatch trends
---
# /agent-fitness
Three modes:
**`/agent-fitness extract`** — runs `python3 daemons/agent-fitness-extract.py` to scan the latest JSONL transcript and append new dispatch rows to `memory/AGENT_FITNESS.md`. Idempotent (dedup by session_id + tool_use_id).
**`/agent-fitness`** (no args) — runs `python3 daemons/agent-fitness-report.py`. Surfaces:
- Total dispatches (all-time, last 30d, last 7d)
- Per-agent calibration table (Total / Clean / Blocked / Errored / Partial / Ratio)
- Per-agent recent outcome trend (last 10 dispatches encoded as `c7 b1 e0 p0`)
- ⚠ Repeat-blocker callouts (any agent with >=2 blocks in window)
- Top 3 failing (agent, task) pairs
**`/agent-fitness --days 7`** — same as above but windowed to last N days. Useful for "is this agent failing more lately?"
## Computation
Calibration ratio per agent: `clean / (total − errored)`. The denominator excludes hard errors (infrastructure/tool failures) and only counts agent-level outcomes.
Trend signal: any agent with >=2 blocks in the window surfaces a repeat-blocker callout. The classifier was hardened in v0.5.1 to catch real Lyra-style block patterns ("BLOCKED:", "are denied", "blocker-found" etc).
## When to run
- After major sessions where multiple sub-agents fired (today S30d had ~7 dispatches)
- Periodically (weekly?) to spot trends
- Before relying on a sub-agent for something heavy ("am I about to dispatch an agent that's been failing?")
## What it does NOT do
- Does not auto-reroute dispatches based on calibration. Surfaces data; principal decides.
- Does not LLM-classify failures — heuristic only (keyword matching on result content).
- Does not pool across sessions for cluster analysis. Future v0.5.x.
- Does not include task-type clustering. All tasks pooled per agent.
## Cross-links
- [[concepts/Somatic v0.5.0 Agent Fitness]] — spec
- [[memory/AGENT_FITNESS]] — the data
- [[concepts/Somatic Roadmap]]
- [[memory/TRUST_DECAY]] — sibling claim → outcome ledger
© wrg32786 · MIT · wrg32786/aigent-os