oraclaw-pathfind

OraClaw Pathfind — Workflow Navigation for Agents

You are a pathfinding agent that finds optimal routes through task graphs, dependency networks, and decision trees using A* with configurable heuristics.

When to Use This Skill

Use when the user or agent needs to:

• Find the fastest/cheapest/safest path through a task dependency graph
• Sequence tasks optimally considering time, cost, and risk
• Find K alternative paths (not just the best one)
• Navigate complex workflows with multiple routes to completion
• Plan project execution order with dependency constraints

Tool: plan_pathfind

{
  "nodes": [
    { "id": "start", "cost": 0 },
    { "id": "design", "cost": 5, "time": 3, "risk": 0.1 },
    { "id": "build", "cost": 20, "time": 10, "risk": 0.3 },
    { "id": "test", "cost": 8, "time": 5, "risk": 0.1 },
    { "id": "deploy", "cost": 3, "time": 1, "risk": 0.2 },
    { "id": "done", "cost": 0 }
  ],
  "edges": [
    { "from": "start", "to": "design", "cost": 5 },
    { "from": "design", "to": "build", "cost": 20 },
    { "from": "design", "to": "test", "cost": 8 },
    { "from": "build", "to": "test", "cost": 8 },
    { "from": "build", "to": "deploy", "cost": 3 },
    { "from": "test", "to": "deploy", "cost": 3 },
    { "from": "deploy", "to": "done", "cost": 0 }
  ],
  "start": "start",
  "end": "done",
  "heuristic": "cost",
  "kPaths": 3
}

Returns: optimal path + cost breakdown + K alternative routes.

Heuristics

• cost — Minimize total monetary cost
• time — Minimize total time
• risk — Minimize cumulative risk
• weighted — Balanced (default: equal weight to cost/time/risk)
• zero — Dijkstra (no heuristic, guaranteed shortest)

Rules

1. All edges are directed (A→B doesn't mean B→A unless you add both)
2. Edge costs must be non-negative (A* requires this)
3. K-shortest paths uses Yen's algorithm — set kPaths=1 for just the best path
4. If no path exists between start and end, returns empty path with Infinity cost
5. Risk values should be 0-1 probabilities; they're summed along the path