completion

A task completion skill for OpenCLAW agents. Maintains a single view of commitments across multiple roles with priority weighting and conflict surfacing.

What it does

• Task CRUD across multiple roles (default: Work and Personal, add your own)
• Role-weighted priorities — when P1s compete across roles, the skill surfaces the conflict instead of silently resolving it
• Daily standup — a short conversation to build your day: carryover, urgency, one thing that matters
• Weekly review — what moved, what's stuck, what needs attention
• Status tracking — tasks flow through backlog → todo → in_progress → blocked → done with full history

Install

Clone into your OpenCLAW workspace skills directory:

git clone https://github.com/geoffscott/todo-skill.git <workspace>/skills/todo

Restart the gateway or ask your agent to refresh skills. Verify with:

openclaw skills list --eligible

Usage

Talk to your agent naturally:

• "I need to finish the API docs by Friday" → captures a task
• "What's on my plate?" → lists tasks grouped by role
• "Let's do a standup" → interactive daily review
• "Weekly review" → retrospective on flow and stuck items
• "The budget proposal is blocked on finance" → status update with context

How it works

The skill stores tasks in a local SQLite database at ~/.openclaw/completion/tasks.db. The database is initialized automatically on first use with default roles. You can add, rename, or reweight roles at any time.

Tasks belong to exactly one role and have a priority (p1/p2/p3). When tasks at the same priority compete across roles, the skill uses role weights to suggest an ordering — but always surfaces the conflict for you to decide.

File structure

SKILL.md            # Skill instructions (read by OpenCLAW)
references/
  schema.sql        # SQLite schema
scripts/
  init_db.py        # Database initialization (idempotent)

License

MIT