Ideas Lifecycle

Ideas Lifecycle

Purpose

Provide fast, read-only access to deferred ideas for planning and decision discussions.

Inputs

• User request that asks to inspect or evaluate existing ideas
• .ai/project/ideas/ideas.yaml
• Optional focus: status, tag, date window, milestone, or requirement

Outputs

• Filtered idea set with concise rationale
• Recommended next action per idea (keep deferred, promote, invalidate, or review)
• Optional handoff to ideas-governance when the user wants persistence

Read Model

1. Read summary first for quick scope.
2. Read only the ideas entries required for the question.
3. Prefer status/date/tag filtering before deep comparison.

Review Heuristics

• Mark as needs-review when review_after exists and is in the past.
• Mark as promotion-candidate when user goals align with trigger_to_start.
• Mark as possible-invalidation when assumptions in summary or why_now_not are no longer true in user-provided context.

Steps

1. Confirm the user intent is read and analysis.
2. Produce a shortlist that directly answers the user scenario.
3. For each idea, include id, status, and one-line recommendation; include optional fields (review_after, linked_function, expected_effect) when present.
4. If the user asks to persist a decision, transfer execution to ideas-governance.

Boundaries

• Do NOT edit .ai/project/ideas/ideas.yaml.
• Do NOT allocate new IDs.
• Do NOT run autonomous periodic checks. Act only on explicit user requests.

Response Template

Use this table when it improves clarity:

Verification

• Source file .ai/project/ideas/ideas.yaml exists and is readable.
• Output rows map to real idea IDs in the source file.
• Recommendations are labeled as suggestions unless the user explicitly requests persistence.