feat(memory): add project and task guidance

AGENTS.md

@@ -70,7 +70,7 @@ Retrieval rule of thumb:
 
 - **Memory is limited** — if you want to remember something, WRITE IT TO A FILE
 - "Mental notes" don't survive session restarts. Files do.
-- When someone says "remember this" → update `memory/YYYY-MM-DD.md` or relevant file
+- When someone says "remember this" → treat that as an explicit instruction to persist it. Capture it in today's daily note, and promote it to `MEMORY.md`, `memory/projects.md`, or `memory/tasks.json` if it is durable, project-level, or actionable.
 - When you learn a lesson → update AGENTS.md, TOOLS.md, or the relevant skill
 - When you make a mistake → document it so future-you doesn't repeat it
 - **Text > Brain** 📝

memory/2026-03-11.md

@@ -37,6 +37,9 @@
 ## Memory taxonomy/docs
 - Added a compact memory taxonomy section to `AGENTS.md` so future sessions know where different kinds of notes belong.
 - Added `memory/README.md` to explain the folder layout, writing rules, and how embeddings should be used: lookup first, source verification second.
+- Added `memory/projects.md` as a lightweight human-readable registry for bigger-picture ongoing efforts.
+- Added `memory/tasks.md` as a human-readable guide for the structured task system in `memory/tasks.json`.
+- Strengthened the explicit `remember this` rule so future sessions know to persist, promote, and route remembered items appropriately.
 
 ## Follow-up intent
 - Be more disciplined about using the daily note as the default inbox and promoting only durable truths into `MEMORY.md`.

memory/README.md

@@ -25,6 +25,16 @@ This folder is the assistant's working memory on disk.
 - `plans/`
   - Multi-session plans or ongoing initiatives
 
+- `projects.md`
+  - Human-readable registry of active bigger-picture efforts
+  - Good for quick orientation across sessions
+
+- `tasks.json`
+  - Structured task state for follow-up and reminders
+
+- `tasks.md`
+  - Human-readable guide for how task tracking should work
+
 - `references/`
   - Stable supporting material worth keeping around for reuse
 
@@ -44,8 +54,10 @@ Embeddings are the lookup layer:
 
 When in doubt:
 1. write to today's daily note
-2. later promote durable truths into `../MEMORY.md`
+2. if it's a durable fact, promote it into `../MEMORY.md`
-3. only create a named note when the topic genuinely needs a long-lived home
+3. if it's a bigger ongoing effort, add/update `projects.md`
+4. if it needs follow-up or status tracking, add/update `tasks.json`
+5. only create a named note when the topic genuinely needs a long-lived home
 
 Create a named note when:
 - the topic will span multiple sessions

memory/projects.md

@@ -0,0 +1,49 @@
+# Projects
+
+Lightweight registry of active multi-session work.
+
+## Status legend
+- active
+- paused
+- blocked
+- done
+- watch
+
+## Projects
+
+### OpenClaw reliability and ops
+- Status: active
+- Goal: keep the local OpenClaw setup stable, safe, and low-friction for daily use
+- Current threads:
+  - improve subagent / ACP reliability
+  - maintain memory hygiene and retrieval quality
+  - keep backups, search, and messaging paths healthy
+- Related memory:
+  - `memory/incidents/`
+  - `memory/ops/`
+  - `memory/tasks.json`
+
+### Daily brief / calendar direction
+- Status: watch
+- Goal: build a useful low-noise daily brief flow, likely calendar-first
+- Current state:
+  - direction chosen on 2026-03-10
+  - no calendar backend configured yet
+  - next decision depends on which calendar backend Will actually uses
+- Related memory:
+  - `memory/2026-03-10.md`
+
+### Flynn council pipeline
+- Status: active
+- Goal: get Flynn's council pipeline working against real models, then make it delegable from zap
+- Current state:
+  - tracked as a task in `memory/tasks.json`
+  - implementation notes live in `memory/plans/flynn-council-fix.md`
+- Related memory:
+  - `memory/plans/flynn-council-fix.md`
+  - `memory/tasks.json`
+
+## Notes
+- Use this file for cross-session project state humans may want to skim quickly.
+- Keep detailed steps/checklists in `memory/plans/` or `memory/tasks.json`.
+- Don’t create a project entry for every tiny task; this is the bigger-picture map.

memory/tasks.md

@@ -0,0 +1,36 @@
+# Tasks
+
+Human-readable companion to `tasks.json`.
+
+## Purpose
+- `tasks.json` is the structured source of truth for active task state.
+- This file explains how to use it and what kinds of items belong there.
+
+## Task statuses
+- `open` → captured, not actively being worked yet
+- `in-progress` → currently being worked
+- `blocked` → waiting on a dependency, decision, or fix
+- `done` → explicitly confirmed complete
+
+## Priority meanings
+- `high` → important / urgent / likely to need follow-up
+- `med` → useful but not urgent
+- `low` → backlog / nice-to-have / revisit later
+
+## When to add a task
+Add a task when:
+- Will asks for something that should be tracked over time
+- there is a follow-up commitment
+- there is a deferred fix or investigation
+- there is a project thread with multiple steps
+
+Do not add a task for:
+- tiny one-off answers
+- trivial actions completed immediately with no follow-up value
+- noisy internal micro-steps
+
+## Operating rules
+- Keep active state in `tasks.json`
+- Append notable updates to the current daily note when useful
+- Mark done only on explicit confirmation or very clear completion evidence
+- Prefer a few meaningful tasks over dozens of tiny chores