io.github.hifriendbot/cogmemai

Delete multiple memories at once by their IDs. Maximum 100 IDs per call. This is permanent.

Update multiple memories at once. Each item needs a memory_id and fields to update. Maximum 50 items per call.

Consolidate related memories into fewer, richer memories. Finds clusters of memories sharing the same subject (3+ memories required), then uses AI to synthesize each cluster into 1-2 comprehensive facts. Originals are archived (not deleted) with full version history. Use dry_run=true to preview without making changes. Great for cleaning up memory clutter after many sessions.

Move a memory to the trash by its ID. Trashed memories are hidden from recall and restorable with restore_memory for 30 days, then purged permanently.

Delete a mandatory rule by its memory ID. Use list_rules to find the ID first.

Export all memories as JSON. Use this to back up memories or transfer them to another project.

Extract memories from a conversation exchange using AI. Send the developer message and assistant response, and the server identifies facts worth remembering (architecture decisions, preferences, bug fixes, etc.).

Analyze memory clusters and extract underlying factual principles or patterns. Principles are observations about your project — "this codebase tends to have X" — not behavioral instructions (those are skills). Use dry_run to preview candidates first.

Signal whether a recalled memory was useful or irrelevant. Helps improve future recall quality over time. Use after recalling memories to indicate which were helpful vs noise.

Manually trigger skill generation from your corrections, preferences, and patterns. Skills are behavioral directives that auto-improve how the AI works with you. CogmemAi also generates skills automatically when enough evidence accumulates — this tool lets you trigger it manually or preview candidates.

Get a memory health dashboard with insights: most recalled memories, never-recalled memories, stale memories, growth trends, and breakdowns by type and category. Use this to identify cleanup opportunities and understand memory usage patterns.

Show what files changed since your last session. Compares the current git state to a snapshot saved when your previous session ended. Helps you understand what happened between sessions.

Read the project's Intent document: the owner's plain-English source of truth (what the project is for, what must always hold, what was decided and why, what is out of scope). It is loaded automatically at session start; call this when you need the full text, or to confirm one exists before proposing a change to it.

View all memories linked to a specific memory. Returns the relationship type and full memory details for each connection. Use this to explore the knowledge graph around a memory.

View the edit history of a memory. Shows all previous versions with timestamps and what changed. Useful for understanding how a decision or fact evolved over time.

Load top memories for the current project plus relevant global memories. Use at the start of a session to get full context from previous sessions. Optionally pass context to get memories most relevant to your current task.

Find memories that may be outdated based on age and access patterns. Returns memories that haven't been recalled or updated recently, so you can review, update, or delete them.

Get tasks for the current project. Returns tasks filtered by status — defaults to showing pending and in_progress tasks. Use at session start to pick up where you left off.

Get current usage statistics — memory count, extractions this month, tier info, projects.

Ask before acting. Judges a proposed action (a command, a message about to be sent, a purchase, a change, anything) against the rules this person has asked their Ai to keep, plus the NEVER and MUST lines of the project intent. Returns allow, ask or deny with the rule that applies. The shell hook does this automatically for commands; call this for everything else, and always before an action that is hard to undo. Fails open: an error is an allow with judged=false.

Bulk import memories from a JSON array. Each memory needs at minimum a content field. Deduplication is applied automatically.

Extract memories from a document by splitting it into chunks and processing each one. Great for onboarding — feed in READMEs, architecture docs, or API specs to quickly build project context.

Connect two related memories with a named relationship. Use this to build a knowledge graph — e.g., linking a bug fix to the architecture decision that caused it, or connecting a preference to the pattern it led to.

List stored memories with optional filters by type, category, scope, or project.

List all mandatory rules for the current project and/or globally. Rules are always-on memories that surface in every session.

List all tags in use across your memories, with counts. Use this to see what threads/groups exist and find related memories by tag.

List trashed (deleted) memories, newest first, with days remaining before each is purged permanently. Restore any of them with restore_memory.

Proactive recall. Call this BEFORE making any suggestion, recommendation, or action plan. Sends the topic to a fast keyword search and returns any relevant prior context: previous attempts, decisions, contacts, evaluations. This prevents suggesting things that were already tried, rejected, or completed. Fast and cheap, use liberally.

Promote a project-scoped memory to global scope so it applies across all projects. Use this when you discover a preference or pattern that should be universal — e.g., "user prefers tabs over spaces" or "always use Bun instead of npm".

Search stored memories using semantic search. Returns memories ranked by relevance, importance, and recency. Use this to find relevant context from past sessions.

Restore a trashed (deleted) memory by its ID. Works within 30 days of deletion, before the trash is purged. Use list_trash to find restorable memories.

Review finished work against the intent: what the person set out to have done, what must always hold, what was decided. Pass a description, output, message or transcript of what you did; the project intent document is used, or pass intent inline when there is none. Returns a plain summary, what the intent covers, what it does not, and any violations. The Stop hook does this automatically for code changes in a repository; call this for everything else.

Save a correction pattern — what went wrong and what the right approach is. These are surfaced automatically when similar situations arise in future sessions, helping avoid repeated mistakes.

Store a developer memory (fact, preference, decision, architecture detail). Memories persist across all Claude Code sessions and are available in future conversations.

Save a mandatory rule that will ALWAYS be followed in every session. Rules bypass scoring and decay — they are injected into every conversation, every time. Use for absolute requirements like "NEVER do X" or "ALWAYS do Y".

Save a summary of the current coding session. Captures what was accomplished, decisions made, and next steps. Stored as a session_summary memory for future reference.

Create a task that persists across sessions. Tasks are tracked with status (pending, in_progress, done, blocked) and priority (high, medium, low). Use this to maintain continuity on multi-session work.

Create or replace the project's Intent document in the owner's own words: a short markdown document with the sections Purpose, Invariants (NEVER and ALWAYS sentences the guard will enforce), Decisions (with the why), and Out of scope. Every session loads it, the end-of-turn review checks changes against it, and every replacement is versioned. Use it when the owner describes what the project is for, or agrees to record a change of intent. Ask before replacing an existing document, and pass changed_by "agent" when you drafted the text.

Set a reminder that surfaces automatically at the start of your next session. Use for follow-ups, things to check, or deferred work. Reminders auto-archive after being shown.

Update an existing memory's content, importance, or scope.

Update a task's status, title, description, or priority. Use this to mark tasks as in_progress, done, or blocked as you work.