Skills

team-harness supports the Agent Skills standard — a cross-tool format for giving AI agents specialized knowledge and workflows.

A skill is a directory containing a SKILL.md file with YAML frontmatter (a name and description) plus markdown instructions. The coordinator sees skill metadata at startup and reads the full instructions on demand via its read_file tool when a task calls for it. Skills are instructions the coordinator reads — not executable code the harness runs on its own.

Skill directories

Project skills override user-global skills of the same name. The .agents/skills/ path matches the Codex CLI convention, so skills written for Codex work in team-harness unchanged.

Creating a skill

mkdir -p .agents/skills/my-skill
cat > .agents/skills/my-skill/SKILL.md << 'EOF'
---
name: my-skill
description: Summarize files and produce a brief report. Use when the user asks for a summary or overview.
---
 
# My Skill
 
## Steps
 
1. Read the target files using `read_file`
2. Summarize the key points
3. Write a brief report
 
## Notes
 
- Keep summaries under 500 words
- Focus on actionable insights
EOF

Naming rules

• 1–64 characters: lowercase letters, digits, and hyphens only.
• Must not start or end with a hyphen, and no consecutive hyphens.
• The directory name is the canonical skill name.

Optional subdirectories

The agent reads these files on demand via its file tools — they are not loaded at startup.

Trust: the harness never executes a SKILL.md — it reads it. But a skill directory can ship a scripts/ folder whose code an agent may run via the unsandboxed bash tool, with the harness process's full privileges. Treat skill directories as trusted, like your PATH. See Troubleshooting → Trust model.