hive-mind

Walk through hive-mind installation and authentication

Use this agent when historical context might help. This agent searches past Claude Code sessions and returns relevant quotes about decisions, preferences, issues, failed approaches, and discussions. **Use when users ask about past sessions**—"do you remember...", "what did we discuss...", "have we tried this before?" **Consider spawning when** planning something where similar work may have been done before in the project, or when a task involves many user preferences (style, conventions, tooling choices). **Can be spawned in parallel with the Explore agent** to get both current code and historical context. When spawning, include: in-context references (things mentioned in conversation), details of planned activities (the agent finds similar past work), and what you're looking for. <example> Context: Continuing work from a previous session user: "Let's keep working on the slides" assistant: "Let me find the current slides and check what we discussed." [Spawns Explore agent: "Find the slides or presentation files the user is working on."] [Spawns retrieval agent: "User wants to continue work on slides/presentation. Looking for: previous discussions about content, style decisions, feedback given, where we left off, any outstanding items."] <commentary> Parallel spawning for continuation: Explore finds current work, retrieval finds context and decisions. Even vague references ("the slides") can be resolved. </commentary> </example> <example> Context: Research iteration user: "Please iterate on the evaluation prompts to see if we can better elicit the behavior" assistant: "Let me explore the current prompts and check what we've tried before." [Spawns Explore agent: "Find the evaluation prompts and understand the current approach to eliciting the target behavior."] [Spawns retrieval agent: "Working on evaluation prompts to elicit behavior. Looking for: past iterations and what worked/didn't, hypotheses about why certain approaches failed, user feedback on previous attempts, any insights about the target behavior."] <commentary> Parallel spawning for research work: Explore finds the current implementation, retrieval finds what's been tried and learned. </commentary> </example>

Retrieval instructions for searching session history. Auto-loaded by the hive-mind:retrieval agent - prefer spawning that agent rather than invoking this skill directly.

1 hook across 1 event