planner

name	planner
description	Create detailed system architecture and 7-day sprint task breakdown for a hackathon project
metadata	{"openclaw":{"always":false}}

Project Planner Skill

Purpose

Transform a selected project idea into a complete technical blueprint: system architecture, tech stack decisions, and a day-by-day implementation plan.

When to Use

After hackathon/selected_idea.json exists.

CRITICAL: Execute Directly — DO NOT use spawn()

Do NOT call spawn() for this task. Run all steps in this single phase execution and write the files directly.

Do NOT run monitoring loops such as:

• exec("sleep ...")
• repeated list_dir(...) polling
• tailing logs/envelopes to wait for subagents

Do NOT call web_search. Use local files as primary truth. web_fetch is optional only if a known URL is strictly needed.

Completion condition is strict: Your task is complete only after both files exist and are non-empty:

• hackathon/plan.md
• hackathon/tasks.json

If context is incomplete, produce a best-effort v1 plan from available inputs and explicitly mark assumptions.

---

Direct Execution Steps

Step 1 — Load Inputs (must do first):

read_file("hackathon/selected_idea.json")
read_file("hackathon/context.json")
list_dir("hackathon/")

Step 2 — Draft Architecture (plan.md) Include all sections below:

• Project objective and MVP boundary
• System components with responsibilities
• End-to-end data flow (ASCII diagram)
• Sponsor integrations with exact roles:
  • FLock: inference path and model call surface
  • Virtual Protocol: on-chain agent identity flow
  • Unibase: persistent memory flow
• Core data models (JSON schema or dataclass-like spec)
• API boundaries (endpoints, input/output shape, auth notes)
• Dependency list with versions (or pinned target versions)
• Risks + mitigations
• Assumptions (explicit)

Step 3 — Build Executable Task Plan (tasks.json) Requirements:

• At least 4 epics
• Each epic has at least 3 tasks
• Every task has concrete deliverable and dependency list
• Priorities use only: critical | high | medium
• Components use only: backend | frontend | ai | blockchain | infra | testing
• Day allocation must be 1..7 and realistic
• Include acceptance-ready tasks for testing and demo prep

Use this JSON shape:

{
  "project_name": "string",
  "architecture_summary": "2-3 sentences",
  "tech_stack": {
    "backend": "Python 3.11 + FastAPI",
    "ai_primary": "FLock API (qwen3-30b-a3b-instruct-2507)",
    "ai_privacy": null,
    "blockchain": "string or null",
    "storage": "string",
    "frontend": "string or null"
  },
  "epics": [
    {
      "id": "E1",
      "name": "string",
      "component": "backend|frontend|ai|blockchain|infra|testing",
      "day_target": 1,
      "tasks": [
        {
          "id": "T1.1",
          "title": "string",
          "description": "what exactly to build",
          "deliverable": "what done looks like",
          "component": "backend|frontend|ai|blockchain|infra|testing",
          "estimated_hours": 2,
          "priority": "critical|high|medium",
          "dependencies": [],
          "day": 1
        }
      ]
    }
  ],
  "risk_register": [
    {"risk": "string", "probability": "high|medium|low", "mitigation": "string"}
  ]
}

Step 4 — Write Files (no delay, no waiting):

write_file("hackathon/plan.md", <markdown plan>)
write_file("hackathon/tasks.json", <valid JSON string>)

Step 5 — Sanity Check and Stop

• Optionally run one quick parse check for JSON validity.
• Do not spawn anything after writing files.
• Do not ask questions; do not wait for further input.

Output Files

• hackathon/plan.md — human-readable architecture doc
• hackathon/tasks.json — machine-readable task list