About Alignment MCP
Alignment MCP is an MCP server published by screen-watcher-elite in the Documentation category: alignment Sentinel MCP server with 7 safety benchmarks and arXiv research tools.
Recent commit activity is the strongest signal that reported issues will get attention.
Install
From source
git clone https://github.com/screen-watcher-elite/Alignment-MCPThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Tools (4)
audit_eval_metricsQuery 7 safety benchmarks (HarmBench, TruthfulQA, StrongREJECT, WMDP, AdvBench, MMLU-Safety, Machiavelli) with O(1) indexed lookups.
Parameters not published.
check_safety_taxonomyLook up 11 alignment concepts (Constitutional AI, RLHF, RLAIF, Red Teaming, Mechanistic Interpretability, Deceptive Alignment, RSP, etc.) with key papers and Anthropic relevance.
Parameters not published.
fetch_paper_breakdownFetch any arXiv paper by ID and get a structured breakdown: core hypothesis, methodology, threat model, limitations, and safety relevance. Cached.
Parameters not published.
search_alignment_papersSearch arXiv for alignment & safety papers with automatic category filtering and 5-minute result caching.
Parameters not published.
Ask AI
✦
Ask AI about Alignment MCP
Powered by Claude · Grounded in docs
I know everything about Alignment MCP. Ask me about installation, configuration, usage, or troubleshooting.
Security checks
6 checks have not been run yet.