20 tools from the Alexandria MCP Server, categorised by risk level.
View the Alexandria policy →alexandria_ask DEFAULT TOOL — Use this for any task that could benefit from specialist expertise. Routes to the best specialist AI agent(s) from Alexandria's libr... alexandria_audit Submit an agent output for causal audit — checks if the output reached the right conclusion through sound reasoning or just got lucky. Returns reas... alexandria_capabilities Discover the full Alexandria platform — not just the 7 MCP tools, but the complete feature surface including tournaments, debates, agent DNA evolut... alexandria_get_tools Export all Alexandria agents as OpenAI-compatible function-calling JSON schemas. Useful for integrating Alexandria agents into non-MCP LLMs that su... alexandria_grade Grade any output against a rubric. Returns per-criterion scores (0.0-1.0) and an overall pass/fail. Use to evaluate quality of agent outputs, code,... alexandria_health Check Alexandria API health and service status. alexandria_learn Store a learning or insight into Alexandria's memory system. This helps agents improve over time by remembering what worked well. Call this after a... alexandria_list_agents List all available specialist agents in Alexandria. Returns agents grouped by team with their names, roles, and descriptions. Use this to discover ... alexandria_outcome Record a real-world outcome for a previously-executed agent task. 48 hours after each execution, Alexandria can prompt you: "did this actually work... alexandria_recommend_learnings Search Alexandria's learnings memory for insights relevant to a query. Uses RAPTOR hierarchical search — returns a mix of specific leaf learnings, ... 2/5 alexandria_spend Report token consumption and USD spent on Alexandria agents. Defaults to a 7-day summary with top-5 agents by cost. Use mode=by-agent for a full pe... alexandria_status Check the status of a long-running Alexandria job (operator run, debate, council, or DNA competition). Use when a previous tool call returned a job... alexandria_tournament View agent tournament rankings and results. Tournaments score agents on standardized tasks — higher ELO means better real-world performance. Read-o... alexandria_council Run the LLM Council — 5 advisor archetypes (contrarian, first-principles, expansionist, outsider, executor) analyze your task in parallel, then a r... 3/5 alexandria_debate Run inter-agent debates for adversarial analysis. Pits agents against each other to stress-test ideas. Modes: adversarial (1v1), panel (1 vs N revi... 3/5 alexandria_doctor Run a self-diagnostic check on the Alexandria platform. Verifies: API connectivity, provider key status, .alexandria/ project detection, agent cata... 3/5 alexandria_operator Run multi-agent goal decomposition. Breaks a high-level goal into 2-6 agent steps and executes them sequentially. Use for complex goals that need m... 3/5 alexandria_premortem Run a pre-mortem failure analysis on a plan or task before execution. Returns failure modes with probability, impact, and mitigations. Use before c... 3/5 alexandria_run_agent Run a specific named Alexandria agent directly on a task. Use this when you know exactly which specialist you want (e.g., "brand-copywriter", "seni... 4/5 The Alexandria MCP server exposes 20 tools across 3 categories: Read, Write, Execute.
Use Intercept, the open-source MCP proxy. Write YAML rules for each tool — rate limits, argument validation, or deny rules — then run Intercept in front of the Alexandria server.
Alexandria tools are categorised as Read (13), Write (1), Execute (6). Each category has a recommended default policy.
Open source. One binary. Zero dependencies.
npx -y @policylayer/intercept