Secure Autonomous Register
Server Details
Read-only, sourced cases of AI agents, vehicles and robots acting on their own, mapped to controls.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-11-25
- URL
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 6 tools
The tools mostly target distinct resources: cases, controls, rules, and signals. There is slight overlap between find_rules and get_control because both surface regulation-to-control mappings, but the resource focus differs enough for an agent to choose correctly.
Most names follow a predictable verb_noun pattern: find_rules, get_case, get_control, list_controls, search_cases. recent_signals is a noun phrase rather than a verb-led name, a minor deviation but still clear.
Six tools is well scoped for a read-only register covering rules, controls, cases, and signals. Each tool has a distinct retrieval role and none feels redundant.
The surface covers the core lifecycle: search and retrieve cases, list and inspect controls, find rules, and monitor new signals. Minor gaps like a dedicated list-all-cases operation are workable through search_cases.
Available Tools
6 toolsfind_rulesFind rulesARead-onlyIdempotentInspect
Regulations and standards that bear on autonomous systems, filtered by keyword, jurisdiction (e.g. 'European Union', 'United States', 'California'), or control ID. Each lists the controls it maps to and how strongly (required, implied, recommended, proposed).
Inputs: limit, query, control, jurisdiction.
get_caseGet a caseARead-onlyIdempotentInspect
The full case: every claim with its type (fact, assertion, interpretation) and source links, the outcome, controls that failed, were absent or held, the Register's labeled interpretation, right-of-reply status for each named party, corrections, and a ready citation.
Inputs: id (required).
get_controlGet a controlARead-onlyIdempotentInspect
One control: its definition, the cases where it failed, was absent or held, and the regulations and standards that require, imply or recommend it.
Inputs: id (required).
list_controlsList controlsARead-onlyIdempotentInspect
The Register's controls (the safeguards that would interrupt the failures it records), each with how many published cases record it as failed, absent or held.
Inputs: no named parameters.
recent_signalsRecent signalsBRead-onlyIdempotentInspect
The Register's nightly stream of new primary records (vulnerability entries, recalls, rules) touching systems that act on their own, newest first. Signals never assign fault.
Inputs: kind, limit.
search_casesSearch casesARead-onlyIdempotentInspect
Find Register cases by keywords (company, system, failure, e.g. 'coding agent deletes database', 'robotaxi', 'prompt injection email'), optionally filtered by world (physical, bridge, software, transaction, boundary) or by a control ID that was missing or failed. Returns short summaries with URLs.
Inputs: limit, query, world, control.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Read-only finance and operations controls for AI agents with evidence and safe next actions.
Human-reviewed real-world actions for AI agents.
Source-linked robotics projects, bills of materials and components for AI agents.
Read-only AI Adoption Controls Framework: controls, Level 3 criteria, standards cross-references
Related MCP Servers
- AlicenseAqualityAmaintenanceAuditable records of human decisions over AI agent work. Approvals, edits, overrides, escalations.39122Apache 2.0
- AlicenseNot gradedqualityBmaintenanceLets AI agents lease a single Windows window or a whole screen, see its controls, click, type and hotkey inside it (in the background, so the user keeps their own mouse and keyboard), and record what they did as a step-by-step tutorial. Every action passes through policy checks — one driver at a time, hard denies on elevated/password/secure-desktop targets, approvals for risky clicks, and logged refusals.MIT
- AlicenseNot gradedqualityBmaintenanceDeterministic Multi-Agent System (MAS) control plane for SRE governance, failure taxonomy diagnosis, and cryptographic WORM trust ledger audit.MIT
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to perform human-authorized, fully auditable GUI actions — clicking, typing, and window operations — on a local Windows desktop, gated by time- and count-limited leases that are denied by default until a person signs off in a popup. Every action is bound to a fresh screen capture, leaves screenshot-based evidence and decision logs on the local machine, and can be halted instantly via a panel button, hotkey, or kill-switch file.2Apache 2.0
Glama MCP Gateway
Add one secure layer between your agents and this server.