Retrieve and inject project-specific entities from the Memstead graph into conversation context.
npx skills add https://github.com/memstead/memstead --skill learn-memstead
Please help me install this Agent Skill. Skill: learn Source: https://github.com/memstead/memstead/tree/main/plugins/claude-code/skills/learn Command: npx skills add https://github.com/memstead/memstead --skill learn-memstead
This skill solves the problem of fragmented project knowledge by ensuring that relevant entities are retrieved and internalized by the AI before it begins any task, preventing the need for re-deriving information.
Ask the agent to learn about the authentication module by typing learn authentication module in your chat interface.
High-intent search queries and answers about installing and using this skill.
To load project knowledge into context, use a retrieval skill to search the Memstead graph for relevant entities and inject the full content directly into the active conversation. This grounds the AI in existing project history before task execution.
Entity retrieval prevents re-deriving information by searching a typed graph structure for validated project knowledge and reading full entity content into the context window. This ensures the agent internalizes existing specifications before starting work.
Pull technical specifications and ADRs into a chat session by executing a contextual knowledge load command like learn authentication module. The agent queries the Memstead MCP server to retrieve schema-validated architectural constraints into the current context.
Yes, you need the Memstead MCP server integration to retrieve conversation context. The server performs validated entity lookups and token-budgeted reading from the typed graph structure to ensure only relevant knowledge is injected.
The best way to ground an AI agent in existing project history is schema-validated retrieval. By searching the typed graph for relevant topics and reading full entity content, the agent receives accurate, validated knowledge instead of guessing.
Yes, entity injection uses token-budgeted reading to manage context limits. The retrieval mechanism searches the Memstead graph and reads full entity content while respecting the active conversation token budget to prevent overflow.
rhpdsConfigure persistent Memory MCP knowledge graph storage for Claude Code sessions.
mdmagnuson-creatorAnalyze unstructured project specifications and output a structured RequirementsManifest in JSON.
memsteadSynchronize project mems with source documentation via validated updates.
ephesGenerate a self-contained reviewer prompt with file paths and review criteria.
