Liminality Mcp — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Liminality Mcp (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
<p align="center"><img src="assets/logo.png" width="120" alt="Liminality"></p>
<h1 align="center">Liminality</h1>
<p align="center">Hand it the hard part: tough questions, real decisions, multi-step tasks, over MCP.</p>
Liminality takes a hard question, a decision, or a multi-step task, breaks it into the sub-questions that actually decide the answer, ties each one to a real tool or endpoint, and gives you back a worked result you can check and reuse. For a choice you get a scored decision frame; for a question, a grounded answer. It earns its keep on the work that one-shot guessing gets wrong, the hard and high-stakes stuff, not quick lookups.
Every solved ask is saved as a reusable route in a shared library, so the second time someone asks something close, it comes back faster and costs less. It runs as a hosted remote server over streamable HTTP. The first few asks on a new key are free.
Remote server, nothing to install:
{ "mcpServers": { "liminality": { "url": "https://mcp.physea.ai/mcp" } } }https://mcp.physea.ai/mcpX-API-Key or Authorization: Bearer) or OAuth2. Free tier on a new key.(decompose → ground to real tools → score the decision) and returns a worked result: a scored decision frame for a choice, or a grounded answer for a question.
research — a deeper multi-source pass that pulls real information for a question.ask_form / apply_form — when the answer depends on your specifics, it hands back a shortmultiple-choice form; relay it, then apply_form folds the answers into a sharper result.
get_my_context — what's known about you and what you've connected.register_asset / set_preference — tell it about your material and how you like results.report_feedback / report_outcome — tell it how a result did, so the routes improve with use.composio_connect — connect a tool so an action can run against it.https://mcp.physea.ai/.well-known/mcphttps://mcp.physea.ai/.well-known/agent.jsonhttps://mcp.physea.ai/.well-known/ai-plugin.jsonhttps://mcp.physea.ai/llms.txt© 2026 Physea. All rights reserved. "Liminality", the name, and the logo are trademarks of Physea. This repository is a listing for a hosted service; no rights to the service, its software, or its data are granted. Use of the service is governed by the terms at https://physea.ai/mcp.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.