harness:status — Claude Code skill: what it does, how to install
Displays evolution progress, scores, and iteration counts. Alerts when optimization stagnates or regresses.
npx skills add raphaelchristi/harness-evolver --skill status -a claude-codeWhat harness:status does
The skill renders the evolution run chart using evolution_chart.py with .evolver.json configuration. It locates utilities in .evolver/tools or $HOME/.evolver/tools. Execution relies on the Python binary at $HOME/.evolver/venv/bin/python or the default python3 on the system path.
It inspects recent scores right after rendering. If the last 3 scores remain within 1% of each other, it warns of stagnation and suggests running /harness:evolve with an architect trigger. Any dip below a previously recorded best score triggers a regression alert.
When LangSmith metadata is present, the skill prints the URL for the top experiment. The command fails immediately if .evolver.json is missing from the project directory.
What it helps with
- display evolution progress and current scores
- detect optimization stagnation across recent iterations
- warn about regressions against previous best scores
- print the LangSmith experiment URL for the best run
Example prompts for the agent
- /harness:status
- Show current evolution scores and how many iterations ran
- Check if the optimization loop is stagnating
How to install harness:status
1Skills CLI (Vercel)
npx skills add raphaelchristi/harness-evolver --skill status -a claude-code2Manually: put the files in place
~/.claude/skills/status/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/status/
1Skills CLI (Vercel)
npx skills add raphaelchristi/harness-evolver --skill status -a codex2Manually: put the files in place
.agents/skills/status/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add raphaelchristi/harness-evolver --skill status -a cursor2Manually: put the files in place
.cursor/skills/status/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install raphaelchristi/harness-evolver status2Manually: put the files in place
.github/skills/status/SKILL.md1Skills CLI (Vercel)
npx skills add raphaelchristi/harness-evolver --skill status -a gemini-cli2Platform CLI
gemini skills install https://github.com/raphaelchristi/harness-evolver --path skills/statusManually: put the files in place
.devin/skills/status/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add raphaelchristi/harness-evolver --skill status -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 harness:status
How do I install harness:status in Claude Code?
Run npx skills add raphaelchristi/harness-evolver --skill status -a claude-code in your terminal.
What does harness:status do?
It runs evolution_chart.py against .evolver.json to plot run history and flag stagnation or score regressions.
What dependencies does harness:status need?
It requires python3 or a virtual environment in $HOME/.evolver/venv, tools in .evolver/tools, and a .evolver.json configuration file in the project root.
Source text
SKILL.md · MIT
---
name: harness:status
description: "Use when the user asks about evolution progress, current scores, best version, how many iterations ran, or whether the loop is stagnating."
allowed-tools: [Read, Bash]
---
# /harness:status
Show current evolution progress.
## What To Do
### Resolve Tool Path
```bash
TOOLS="${EVOLVER_TOOLS:-$([ -d ".evolver/tools" ] && echo ".evolver/tools" || echo "$HOME/.evolver/tools")}"
EVOLVER_PY="${EVOLVER_PY:-$([ -f "$HOME/.evolver/venv/bin/python" ] && echo "$HOME/.evolver/venv/bin/python" || echo "python3")}"
```
### Display Chart
```bash
$EVOLVER_PY $TOOLS/evolution_chart.py --config .evolver.json
```
### Additional Analysis
After displaying the chart:
- Detect stagnation: if last 3 scores within 1% of each other, warn and suggest `/harness:evolve` with architect trigger.
- Detect regression: if current best is lower than a previous best, warn.
- Print LangSmith experiment URL for the best experiment if available.
© raphaelchristi · MIT · raphaelchristi/harness-evolver
