From agentdb-memory
Retrieve relevant memories for the current task from AgentDB. Use at the start of a task to load prior knowledge, when stuck to surface what worked before, or when the user asks "what do we know about X" / "have we done this before?"
How this skill is triggered — by the user, by Claude, or both
Slash command
/agentdb-memory:agentdb-recallThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Pull relevant past memories into the current context.
Pull relevant past memories into the current context.
| Need | Tool | Returns |
|---|---|---|
| Similar past tasks | agentdb_reflexion_recall | Episodes, top-k by similarity |
| Lessons from past failures | agentdb_critique_summary | Combined critique text |
| What worked last time | agentdb_success_strategies | Approach summaries from high-reward episodes |
| Generic patterns / facts | agentdb_pattern_search | Patterns ranked by similarity |
| Reusable skills by intent | agentdb_skill_search | Skills ranked by precondition match |
agentdb_reflexion_recall with k=5, minReward=0.5.agentdb_pattern_search with the same query, k=5.agentdb_record_feedback (or recordFeedback via the library) so the bandit learns.minReward — drop low-quality matches (default 0.3).onlyFailures — explicitly query the postmortem set when debugging.onlySuccesses — only winning approaches when copying a strategy.timeWindowDays — recent context only when the codebase has shifted.npx claudepluginhub ruvnet/agentdb --plugin agentdb-memoryStore a memory in AgentDB — an episode (task + outcome + critique), a pattern, or a skill. Use when the user says "remember this", "save this for later", "add to memory", or when the agent has just succeeded/failed at a task and the lesson is worth keeping.
Searches agentmemory for past observations, sessions, and learnings matching a query. Groups results by session, shows type/title/narrative, highlights importance >=7. Triggers on 'recall', 'remember', or prior context needs.
Searches and surfaces relevant memories from past sessions to inform current work with decisions, patterns, and learnings. Supports hybrid, vector, and text search modes with namespace filtering.