mubit-ai

mubit-ai/claude-plugins

Plugin Claude Code

Official Mubit marketplace: persistent, typed, self-improving memory for Claude Code.

mubit-ai/claude-plugins

Plugin Claude Code

Persistent, typed, self-improving memory for Claude Code, backed by Mubit. Captures work involuntarily, recalls relevant lessons before every prompt, and attributes outcomes so memory improves with use.

mubit-ai/claude-plugins

MCP server Claude CodeCodexCursor +2

MCP server "mubit" as configured in mubit-ai/claude-plugins. Runs ${CLAUDE_PLUGIN_ROOT}/mcp/dist/index.js with node.

mubit-ai/claude-plugins

Agent

Deep memory search across several angles, run in an isolated context so the main conversation does not absorb the raw evidence. Use for "what do we know about X" questions needing more than one query.

mubit-ai/claude-plugins

Hook

Runs when a session starts on startup, resume, clear, compact and fork, running node. From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Hook

Runs on the CwdChanged event, running node. From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Hook

Runs when you submit a prompt, before the agent sees it, running node (2 commands). From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Hook

Runs before the agent uses a tool for Bash tool calls, running node (2 commands). From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Hook

Runs when a subagent starts, running node. From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Hook

Runs after a tool call finishes, running node. From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Hook

Runs when a tool call fails, running node. From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Hook

Runs when the agent finishes a response, running node. From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Hook

Runs on the StopFailure event, running node. From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Hook

Runs when a subagent finishes, running node. From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Hook

Runs before the context is compacted, running node. From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Hook

Runs after the context is compacted, running node. From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Hook

Runs when a session ends, running node. From mubit-ai/claude-plugins.

mubit-ai/claude-plugins

Skill Claude CodeCodex

Answer the audit question: list what this Mubit instance actually holds for a run, filter it by time, type, agent or origin, and export the whole record as JSONL you can keep. Read-only.

mubit-ai/claude-plugins

Skill Claude CodeCodex

Sign in to Mubit and store an API key for this machine. Use on a fresh install, when the plugin reports authfailed, or when a key has been rotated or revoked.

mubit-ai/claude-plugins

Skill Claude CodeCodex

Save a named snapshot of where this run has got to, stored verbatim. Use when the user is about to do something risky or destructive, when a compaction is coming that they want to survive on their own terms, or when they ask to mark a point they may need to come back to.

mubit-ai/claude-plugins

Skill Claude CodeCodex

Open the local Mubit dashboard: browse and search lessons, see what recall cost per prompt, and check ingest health. Loopback only, on a port nobody else can reach.

mubit-ai/claude-plugins

Skill Claude CodeCodex

Diagnose Mubit connectivity, memory health, and stuck ingest jobs. Use when memory looks empty, captures are not landing, or the status line shows a failure glyph.

mubit-ai/claude-plugins

Skill Claude CodeCodex

Delete a lesson from Mubit memory, or mark an entry superseded. Prefer a negative outcome for a lesson that is merely wrong — deletion cannot be undone.

mubit-ai/claude-plugins

Skill Claude CodeCodex

Report what is actually in Mubit memory — entry counts, staleness, contradictions and per-section health. Use when memory looks empty or stale and you need to know whether anything was ever stored, rather than whether the instance is reachable.