Memory & Context

Import a wiki graph or OKF bundle into the current Obsidian vault

Ingest documents, folders and URLs into an Obsidian wiki as linked notes

Audit wiki health for broken links, orphans, contradictions and stale pages

Answer questions from your compiled Obsidian wiki with cited, multi-hop reasoning

Archive and rebuild wiki knowledge from source or restore a prior archive

Review staged wiki writes and promote accepted pages to final locations

Report wiki ingestion gaps, pending source changes and maintenance status

Create cross-cutting synthesis pages where wiki concepts recur together

Distill work from the current project into your Obsidian wiki as knowledge

Clear the context of CCB managed agents without deleting project state

Compact CCB managed agent contexts without restarting or losing project state

Decide when to hand off, resume or sprint a long session to manage context

Recover from an irreversible mistake with damage control, repair and lessons learned

Choose the right memory search entry point and control noise when recalling context

Pick the right memory search scope for source tracing, absence checks and deltas

Decide whether a person-related memory clue should become a rejectable proposal

Summarize and compact long agent sessions before context limits cut off important details

Wire memory middleware into TanStack AI chats with recall and save adapters

Add Hindsight hosted memory with retain, recall and reflect tools in TanStack AI

Add Honcho hosted memory to TanStack AI chats with dialectic recall

Set up in-memory chat memory for TanStack AI dev servers, tests and demos

Connect TanStack AI chats to a mem0 server over HTTP for memory recall

Store TanStack AI chat memory in Redis with ioredis or node-redis clients

Decide whether a semantic review packet holds a personal memory worth saving

Keep long agent sessions on track, break loops and hand off before context fills

Clone and refresh remote git repos in a local cache for reference lookups

Read shared pi session transcripts from links and extract the conversation

Write compact, factual project context for handing work to another session

Extract recurring failure patterns from agent logs into reviewable rule proposals

Keep decisions.md and patterns.md current and search memory across tools

Send messages and handoffs to other agent sessions across tools

Set up continual learning, hooks and memory scoping for AI coding agents

Summarize what changed in the chat and list items waiting on the user

Send bulk output to subagents and keep only summaries in the main context

Rebuild recent work context from chat history and shared records into a brief

Manage saved project context, archive stale notes and reconcile working memory

Audit agent memory for hallucinations, circular confirmations and poisoned state

Search your codebase with an AST-based index that agents can query

Save critical files and task state before compaction and restore them after

Manage context with write, select, compress and isolate for longer sessions

Cut token use and fix degraded context when agent sessions slow down

Track coding session costs, set spending alerts and cut token spend

Show session analytics, correction trends and productivity stats from project memory

Save a mistake and its correction as a rule future sessions will reuse

Pull past corrections and lessons into view before starting a new task

Write a structured handoff note with progress, open tasks and decisions

Reduce token waste with tool-call budgets, one-pass coding and read-before-write rules

Wrap up a coding session by auditing changes, running checks and saving learnings

Sync messages from WhatsApp, Discord, Slack and X into daily memory digests