Learning Tools API

Complete reference for the learning tools that detect patterns and generate skills.

suggest_skill

Get skill suggestions for current context.

Parameter	Type	Description
context	string	Current task context
domain	string	Filter by domain
task_type	string	Filter by task type
limit	number	Max suggestions

Response

{
  "suggestions": [
    {
      "skill_id": "skill-abc",
      "name": "CORS Configuration",
      "relevance": 0.92,
      "instructions": "..."
    }
  ]
}

generate_skill

Generate a new skill from memories. Optionally check for conflicts with existing skills.

Parameter	Type	Required	Description
memory_ids	string[]	Yes	Source memory IDs
name	string	Yes	Skill name
domain	string[]	Yes	Target domains
description	string	No	Skill description
check_conflicts	boolean	No	Check for similar/conflicting skills

Response

{
  "skill": {
    "id": "skill-new123",
    "name": "Generated Skill Name",
    "status": "draft",
    "instructions": "... synthesized from memories ..."
  }
}

With Conflict Checking

When check_conflicts: true:

{
  "skill": { ... },
  "conflicts": [
    {
      "skill_id": "skill-existing",
      "name": "Similar Existing Skill",
      "overlap": 0.75
    }
  ],
  "warnings": ["Similar skill already exists in domain"]
}

analyze_patterns

Analyze memories to detect patterns.

Parameter	Type	Description
domain	string	Filter by domain
days	number	Look back period
limit	number	Max patterns

Response

{
  "patterns": [
    {
      "description": "CORS debugging approach",
      "memory_count": 4,
      "confidence": 0.88,
      "memory_ids": ["mem-1", "mem-2", "mem-3", "mem-4"],
      "suggested_skill_name": "CORS Debugging Workflow"
    }
  ]
}

report_outcome

Log outcome of applying a skill.

Parameter	Type	Required	Description
skill_id	string	Yes	Skill that was applied
outcome	string	Yes	`success` or `failure`
notes	string	No	Additional context

Response

{
  "success": true,
  "message": "Outcome recorded",
  "skill_stats": {
    "total_uses": 10,
    "success_rate": 0.9
  }
}

get_context

Get relevant memories and skills for current session.

Parameter	Type	Description
domain	string	Filter by domain
query	string	Context description
memory_limit	number	Max memories
skill_limit	number	Max skills

Response

{
  "memories": [
    {
      "id": "mem-123",
      "observation": "...",
      "relevance": 0.85
    }
  ],
  "skills": [
    {
      "id": "skill-456",
      "name": "...",
      "relevance": 0.92,
      "instructions": "..."
    }
  ]
}

Usage Workflow

1. Work on tasks, storing memories
     ↓
2. analyze_patterns to find repeated approaches
     ↓
3. generate_skill from pattern memories
     ↓
4. Review and approve_skill
     ↓
5. suggest_skill in future similar contexts
     ↓
6. report_outcome to track effectiveness