From vision
Scores OKR key results against predefined thresholds (commit, target, stretch) using a deterministic formula. Called by op-quarterly-planning and op-annual-review to produce reproducible retrospective scores.
How this skill is triggered — by the user, by Claude, or both
Slash command
/vision:flow-okr-scoring-rubricThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
**Trigger:** Called by `op-quarterly-planning` (at the close-out step), `op-annual-review`. User-facing for "score my quarter."
Trigger: Called by op-quarterly-planning (at the close-out step), op-annual-review. User-facing for "score my quarter."
Applies a standard rubric to every key result on a closing OKR set. Replaces "I think we hit ~70% on that one" with a deterministic grade computed from the KR's own success thresholds.
Threshold convention (set at OKR creation, not retrospect):
target — the planned end value (e.g., 50,000).commit — the floor commitment, scoring 0.7 if reached (e.g., 35,000).stretch — the stretch target, scoring 1.0 if reached (e.g., 65,000).Score formula:
KRs without thresholds are flagged and skipped from the score (and surfaced to open-loops so the next OKR cycle sets thresholds up front).
vault/vision/00_current/.commit, target, stretch from the OKR file's frontmatter or inline annotations.01_prior/QYYYY-okrs-scored.md.# QYYYY OKRs — Scored
Overall: [N] / 1.00
## [Objective 1] — [average]
- KR1.1: [final] vs commit [c] / target [t] / stretch [s] → [score]
- KR1.2: ...
## [Objective 2] — [average]
...
## Calibration flags
- [KR description] — missing thresholds
vault/vision/config.md:
okr_threshold_default_commit_pct (default 70 — used to seed when user doesn't specify)okr_threshold_default_stretch_pct (default 130)~/Documents/aireadylife/vault/vision/00_current/QYYYY-okrs.md~/Documents/aireadylife/vault/vision/01_prior/QYYYY-okrs-scored.md, ~/Documents/aireadylife/vault/vision/open-loops.mdnpx claudepluginhub fru-dev3/ai-ready-life --plugin visionCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.