Server facts
ejentum-mcp
Project
- Author
- ejentum
- Repository
- github.com/ejentum/ejentum-mcp
- License
- Not specified
- Category
- AI Models & Agents
Tech
- Runtime / language
- Not specified
- Transports
- stdio
- Remote available
- No
- Authentication
- Not specified
Activity
- GitHub stars
- 16
- Last commit
- Data as of
Tags
- Claude
- agentic ai
- anthropic
- anti deception
- ejentum
- mcp
- model context protocol
- reasoning harness
Overview
Ejentum-mcp connects agent workflows to the Ejentum Logic API for structured guidance before an answer is produced. Its reasoning tools address planning, diagnosis, and analytical decisions; its code tools cover generation, review, debugging, and refactoring. Anti-deception guidance examines pressure to agree, unsupported assertions, and compromised framing. The memory harness instead examines an observation the agent has already made about conversational shifts or recurring behavior. It is intended to refine perception, not store records or retrieve structured data. Adaptive variants tailor the selected procedures and reasoning graphs to the particular task or observation.
A caller supplies a brief query describing the task, integrity concern, or observed signal. The returned text provides procedures and checks for the model to apply internally rather than reproduce as its answer. Supported clients include Claude Desktop, Cursor, Windsurf, Claude Code, and n8n through MCP. The wrapper uses stdio and requires an Ejentum API key. Directly requesting a harness can matter: the documentation notes that agents may skip optional tool calls when they judge their own reasoning sufficient.
ejentum-mcp setup notes
- ejentum-mcp records stdio only; confirm a network transport upstream before planning remote access.
- ejentum-mcp has no recorded license; confirm reuse terms in its repository.