Skill Tools

Tools for managing and executing skills — modular capability extensions for the agent.

execute_skill

Execute a skill by name.

Parameter	Type	Description
skill_name	string	Name of the skill to execute
params	object	Skill-specific parameters (optional)

Classification: Depends on skill type

Return shape:

{
  "status": "success",
  "stdout": "skill execution output",
  "exit_code": 0,
  "duration_ms": 5000
}

Notes:

• Only enabled skills can be executed (controlled via enabled_skills in settings)
• Progressive disclosure: Full skill instructions are injected only when the agent activates the skill

load_skill

Load a skill's full instructions into the agent's context.

Parameter	Type	Description
skill_name	string	Name of the skill to load

Classification: Host

Notes:

• At startup, only skill metadata (name, description) is loaded
• load_skill retrieves the full SKILL.md content and any associated scripts
• This two-phase approach keeps the agent's initial context lean

create_skill

Create a new skill with a SKILL.md template.

Parameter	Type	Description
name	string	Skill name (lowercase, alphanumeric, hyphens)
description	string	What the skill does

Classification: Host

Notes:

• Creates ~/.contop/skills/{name}/SKILL.md with YAML frontmatter template
• Name validation: ^[a-z0-9][a-z0-9\-]{0,63}$

edit_skill

Edit an existing skill's SKILL.md content.

Parameter	Type	Description
skill_name	string	Name of the skill to edit
content	string	New SKILL.md content

Classification: Host

Notes:

• Validates YAML frontmatter before saving
• Frontmatter must include name and description fields

Skill Lifecycle

1. Discovery — Skills are discovered in ~/.contop/skills/ at startup
2. Metadata loading — Only name, description, version, type loaded initially
3. Enable — User enables via settings or API
4. Full load — Agent calls load_skill to get complete instructions
5. Execute — Agent calls execute_skill to run the skill

---

Related: Skills Engine · Tool Layers