journal/Journal.AI/Evening-Review.txt
Jacob Schmidt 192e6e3891 feat: add AI coaching, conversation persistence, and LLamaSharp integration
- Add Journal.AI project with LLamaSharp-based AI service (Phi-3 model)
- Implement coach sessions (daily check-in, evening review, weekly review)
- Add conversation CRUD with SQLCipher persistence
- AI chat with full conversation history for context-aware replies
- Frontend: CoachPanel, AI stores, conversation stores, side panel UI
- Conversation list with create, rename, and delete support
- Fix Phi-3 output quality (system prompt leaking, token cleanup, JSON filtering)
- Fix CREATEDRAFT kind override in coach sessions

Co-Authored-By: Oz <oz-agent@warp.dev>
2026-03-01 16:07:59 -06:00

19 lines
446 B
Plaintext

{{CoachRules}}
Task:
Generate an EVENING REVIEW for date {{dateLocal}}.
Context (JSON):
{{contextJson}}
Preferences (JSON):
{{preferencesJson}}
Output:
Return a CoachPlanDto JSON with kind="evening_review".
Include:
- Brief recap of today
- 2-4 observations with evidence
- 1-{{maxQuestions}} reflection questions
- 1-{{maxNextActions}} next actions for tomorrow
- patchProposal that creates: (optional) tomorrow-intent fragment AND/OR todos.