Skill-Flow-Orchestrator (SFO)

Production-ready Skill Orchestration System that transforms isolated skills into an intelligent, self-coordinating ecosystem.

Core Capabilities

1. Intelligent Routing - Match task DNA to skill DNA for optimal selection
2. Flow Composition - Build skill chains as Directed Acyclic Graphs (DAGs)
3. Context Management - Persist state across skill executions (hybrid: memory file + in-prompt)
4. Chain Templates - Save successful chains as reusable workflows
5. Adaptive Learning - Optimize routing based on execution outcomes

When to Use

• Complex tasks: Require multiple skills working together
• Automated selection: Need intelligent skill routing based on task analysis
• Workflow reuse: Save and replay successful skill combinations
• Context continuity: Need state persistence across skill boundaries
• Pipeline automation: Building multi-stage processing workflows

4 Orchestration Modes

Quick Start

1. Scan Available Skills

python3 .agent/skills/skill-flow-orchestrator/scripts/scan_skills.py

Output: .agent/sfo/skill_index.json - indexed skills with DNA

2. Match Skills to Task

python3 .agent/skills/skill-flow-orchestrator/scripts/match_skills.py "<task_description>"

Returns ranked skills with match scores

3. Orchestrate Workflow

python3 .agent/skills/skill-flow-orchestrator/scripts/orchestrate.py --mode standard "<task>"

Orchestration Workflow

1. INTAKE - Parse task, extract requirements
2. ANALYZE - Match task DNA with skill DNA using semantic matching
3. COMPOSE - Build execution DAG from matched skills
4. EXECUTE - Run skills in order, passing context
5. ADAPT - Log outcomes, update routing weights

Context Management

• Long-term state: .agent/sfo/context.json
• Chain templates: .agent/sfo/templates/
• Execution logs: .agent/sfo/logs/

References

• Orchestration Modes - Detailed mode configurations
• DAG Composition - Building skill chains
• Template System - Saving/loading workflows