Close out the working ISO week: aggregate ops/activity.jsonl events, regenerate the activity table between BEGIN/END sentinels, draft Friday review (wins, misses, lessons to graduate, seed for next week), freeze the week file. Trigger: 'close week', 'Friday review', '/close-week'.
Wrap up the current coding session: update the project wiki page with lessons and decisions, graduate cross-cutting findings to domain / pattern / lesson pages, check git status, present a status table. Trigger: 'end session', 'wrap up', 'done for today'.
Ingest a source document into the wiki following the Karpathy LLM Wiki pattern. Compile content into structured frontmatter + Markdown pages, resolve [[wikilinks]] against the existing graph, update related pages, run lint, emit activity event. Trigger: 'ingest [source]', 'mass ingest [batch]'.
Run wiki health checks (structural + semantic) and report findings by severity: critical (orphans, missing frontmatter), important (cross-reference gaps, thin pages), informational. Trigger: 'lint', 'health check', 'check wiki'.
Plan the upcoming ISO week: resolve active week (TZ Europe/Paris), lift carry-over from prior week's 'Seed for W+1' block, aggregate open - [ ] todos from project + meeting pages, draft a Monday plan grouped by topic. Trigger: 'plan week', 'Monday plan', '/plan-week'.
Produce a 30-second situational-awareness briefing for any child project: tech stack, folder structure, git history, recent changes, deployment status, current state, link to wiki project page. Trigger: 'project status', 'status briefing', '/project-status'.
Answer a question by searching the wiki: read index.md, find relevant pages, synthesize an answer with [[wikilink]] citations, file the answer in wiki/queries/, update log.md. Trigger: 'query [question]' or a direct knowledge-base question.
Instructions for ozgurkarahan/ai-agent-memory, covering resolve the memory root, phase 0: inventory (mass ingest only), phase 1: gather context, phase 2: llm compilation (this is you — use your reasoning) and 2a. what type of knowledge is this?.
Instructions for ozgurkarahan/ai-agent-memory, covering lint wiki — agent-native health check, resolve the memory root, procedure, critical and important.
Instructions for ozgurkarahan/ai-agent-memory, covering new engagement, resolve required roots, step 1: parse the request, step 2: scaffold each project and 2a. create project directory.
Instructions for ozgurkarahan/ai-agent-memory, covering project status, instructions, resolve the memory root, 1. codebase analysis and 2. project history — what we've done.