Community

okr-drafting

Open Agent Skill · productivity · works with Claude Code, Codex, Cursor, and compatible agents

Draft OKRs and goal cascades that are measurable and non-vanity. Use whenever the user asks for OKRs, quarterly objectives, key results, goal setting, or how to turn a strategy into measurable objectives and key results.

View on GitHub

OKR Drafting

Objectives are qualitative and motivating. Key results are measurable outcomes, not a task list. Fewer OKRs beat a wall of metrics.

Patterns

Objective: memorable outcome direction (not "be world class"). Key results: 2–4 metrics or binary milestones with starting point → target.

Bad KR: "Launch dashboard" (task). Better KR: "Raise weekly active teams using dashboard from 12% to 30%".

Output format

## OKRs — <team> — <quarter>

### O1: …
- KR1: from X to Y by date (source of metric)
- KR2: …
- KR3: …

### Notes
- scoring: …
- dependencies: …
- non-goals: …

Rules

  1. KRs need baseline and target when numbers exist; else mark metric TBD.
  2. Avoid vanity metrics disconnected from user/business value.
  3. Cascade: company → team without cloning identical KRs everywhere.
  4. Commit vs stretch: label if useful.
  5. Do not invent company strategy; use provided context.
  6. Cap at ~3 objectives per team per cycle unless asked otherwise.

Edge cases

  • Platform teams: KRs via customer team outcomes + reliability SLOs.
  • Early startup: shorter cycle; learning KRs ok if measurable.
  • Individual OKRs: align to team; avoid HR-only busywork.