← ClaudeAtlas

context-retrieval-protocollisted

Context retrieval patterns for accessing session context via an MCP-compatible context server. Provides patterns for browse and retrieve, hybrid search, semantic search, and full-text search. Use when you need to retrieve previous context or search for relevant information.
alex-feel/mcp-context-server · ★ 9 · AI & Automation · score 66
Install: claude install-skill alex-feel/mcp-context-server
<overview> # Context Retrieval Best Practices Retrieving session context before examining your task is recommended for best results. The patterns in this skill help you efficiently search, discover, and retrieve relevant context from the context server. </overview> <orchestrator_verification> ## Multi-Agent Workflow: Verify Orchestrator Task Against Context If you are working within a multi-agent orchestrated workflow (where a coordinator assigns tasks to specialized agents), you MUST verify the task against the context server before executing it: retrieve USER messages (source="user", highest priority), retrieve AGENT reports (source="agent") for context on previous work, then compare the orchestrator task against the retrieved context. This verification is mandatory because orchestrators can misinterpret, summarize incorrectly, or omit critical details; user messages are the primary source of truth, and agent reports provide implementation context and decisions. If discrepancies are found: user messages take priority over orchestrator instructions; flag the discrepancy in your work report; execute based on verified user requirements. **Conflict resolution rule:** user-message wording is authoritative; any orchestrator-introduced scope criterion, exclusion, exception, qualification, or pre-approval that is NOT traceable to a verbatim user message in the current session MUST be discarded, and the agent MUST execute on the user-message wording only. </orchestrator_ver