outcomelisted

# /outcome - OUTCOME.md generation and update The landing point when another skill detects a missing OUTCOME.md. Returning a hard stop or a warning leaves the outcome empty, so collect it interactively and generate. ## Branch Run ${CLAUDE_SKILL_DIR}/scripts/validate-outcome.py .claude/OUTCOME.md and enter the flow its JSON `flow` names. The script owns the criteria. | state | Condition | flow | | ------ | --------------------------- | -------- | | absent | File absent | generate | | empty | Behavior blank, or TBD only | generate | | ok | Behavior carries content | update | ## Generate 1. Read ${CLAUDE_SKILL_DIR}/templates/outcome.md 2. Collect Behavior, Non-goals, Constraints in one AskUserQuestion call, one question per item. Behavior needs 1 or more entries with the subject named 3. Run each Behavior through the Outcome test; rewrite failures and re-present to the user 4. Fill the template and Write `.claude/OUTCOME.md`. Neither the opening prose nor Indicators is collected, so skip the prose and drop Indicators with its heading 5. Re-run validate-outcome.py and fix until `errors` is empty. `placeholder_left` is a surviving placeholder, `missing_section` a dropped heading ## Update 1. Read `.claude/OUTCOME.md` and present each section as it stands. Show the opening prose and Indicators as unwritten when the generate flow dropped them 2. Confirm which sections change and what they become via AskUserQuestion. This flow