Brain:update Skill for Claude

/brain:update --- End-of-Session Brain Flush

This is a forcing function. When invoked, Claude MUST thoroughly review the entire conversation and write everything learned to the brain DB. No shortcuts, no skipping.

Procedure

Step 1: Detect brain DB

Look for project_brain.db in the current working directory or parent dirs. If not found, error: "No brain DB found. Run /brain init first."

Step 2: Read current DB state (cross-session awareness)

CRITICAL --- the user runs multiple terminal sessions in parallel. Another session may have already written to the brain since this conversation started. Before extracting, read the current DB state:

python3 ~/.claude/skills/brain/scripts/brain/brain_cli.py context {slug}

For each project, note:

• Existing entities (names and types) --- so you don't propose adding what's already there
• Existing decisions (decision text) --- so you don't duplicate
• Existing events (title + date) --- so you don't duplicate
• Current task statuses --- so you don't regress a task that another session already advanced (e.g., don't set to "open" if another session set it to "done")

Conflict rules:

• Entity already exists with newer metadata --- skip your update. The other session's data is fresher.
• Decision already recorded --- skip. Match on first 40 chars of decision text.
• Event already recorded --- skip. Match on title + date.
• Task status conflict --- the MORE ADVANCED status wins. Status order: open < in_progress < blocked < waiting < done < cancelled. Never regress.
• Task notes --- if brain has notes that your session doesn't know about, MERGE (append), don't overwrite.
• New data from this session not in DB --- these are the writes to propose.

Show any conflicts detected:

CROSS-SESSION CONFLICTS
=======================
  Task "Get sandbox access" --- brain says "done" (updated by another session), 
  this session would set "open". KEEPING "done".
  
  Decision "Use Stakeholder role..." --- already in brain. SKIPPING.

Step 3: Scan the full conversation

Go through every message from top to bottom. For EACH project in the brain DB, extract:

| Category | What to look for |

|----------|-----------------|

| New entities | Any person, team, role, system, module, or org mentioned for the first time |

| New relationships | Connections discovered: X owns Y, A reports to B, team scoped to module |

| New decisions | Anything decided, agreed, confirmed, resolved, or ruled out |

| New events | Meetings, calls, emails read, milestones, deployments |

| New tasks | Action items, to-dos, next steps, follow-ups, blockers identified |

| Task updates | Existing tasks that changed status (started, completed, blocked, cancelled) |

| Entity updates | New info about existing entities (role change, new metadata, correction) |

| Thread updates | New items to link to existing threads, or new threads to create |

Step 3: Execute writes immediately

The user invoked /brain-update --- that IS the confirmation. Do NOT ask "Write all to brain?" or wait for approval. Just do it.

Use Python to write all items via the brain operations module:

import sys
sys.path.insert(0, '$HOME/.claude/skills/brain/scripts')
from brain.schema import get_db
from brain.operations import *

Write in this order:

• Entities (upsert --- safe to run repeatedly)
• Relationships
• Tasks (new + updates)
• Threads + thread_items
• Decisions
• Events

Step 6: Report

BRAIN UPDATED
=============
Entities:      +3 (total: 47)
Relationships: +1 (total: 37)
Tasks:         +2, ~3 updated (total: 8)
Decisions:     +1 (total: 4)
Events:        +1 (total: 3)
Threads:       1 updated (total: 1)
Synced to:     MemPalace (N drawers), brain.json (N people)

Step 6b: Sync to MemPalace and brain.json

After all brain DB writes complete, run the memory bridge to sync new knowledge to the other memory stores:

from brain.memory_bridge import full_sync
full_sync(db_path, project_slug)

This:

• Copies new entities, decisions, and events to MemPalace as searchable drawers (wing = project slug, room = entity type)
• Rebuilds brain.json people section from all brain DB person entities across all projects
• Is best-effort — if MemPalace or brain.json sync fails, brain DB data is still safe

The sync is automatic. No user confirmation needed for this step.

Step 7: Update knowledge articles

After brain DB writes complete and full_sync runs (Step 6b), incrementally update

knowledge articles for entities whose evidence has changed in this session.

Collect changed entity IDs from the writes performed in Step 3 (entities upserted,

decisions/events written). Pass these to the knowledge engine for targeted refresh:

import sys, os
sys.path.insert(0, os.path.expanduser("~/.coco/knowledge"))
from engine import KnowledgeEngine

engine = KnowledgeEngine()
# changed_entity_ids = list of brain DB entity IDs written in this session
engine.incremental_update(project_slug, changed_entity_ids)

Or equivalently via CLI (re-harvests all entities but only regenerates stale ones):

python3 ~/.coco/knowledge/cron.py --run --project {slug} --phases 2,3,5

Only articles whose source_hash changed are regenerated. Expected: 1–5 article

regenerations per typical session. Estimated cost: ~$0.01.

Show result inline in the existing brain-update summary block:

BRAIN UPDATED
=============
Entities:      +3 (total: 47)
Relationships: +1 (total: 37)
Tasks:         +2, ~3 updated (total: 8)
Decisions:     +1 (total: 4)
Events:        +1 (total: 3)
Threads:       1 updated (total: 1)
Synced to:     MemPalace (N drawers), brain.json (N people)
Knowledge:     N articles refreshed  (N new entities, N updated)

Skip silently if:

• ~/.coco/knowledge/ does not exist (knowledge engine not installed)
• cron.py is unavailable or returns a non-zero exit code
• The knowledge engine step takes more than 60 seconds (non-blocking)

Important Rules

• Be thorough. Scan EVERY message. If someone was mentioned by name, they should be an entity. If something was decided, it should be a decision. If there's a next step, it should be a task.
• Deduplicate. Check what already exists in the DB before proposing new writes. Use upsert_entity which matches on (project_id, type, name) or (project_id, type, external_id).
• Date everything. Decisions and events need dates. Use today's date if the exact date isn't clear.
• Attribute decisions. Record who decided, not just what was decided.
• Link to threads. If a task, decision, or event belongs to an existing thread, link it.