Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →lawve-ai/runtime-admissibility-review
Determines whether a specific AI-agent action, output, recommendation, or proposed commitment remains admissible for execution or institutional reliance under current authority, delegated scope, evidence, facts, policy, risk, escalation, and revocation conditions. Use this Skill before an enterprise or regulated AI agent executes, updates records, triggers workflows, communicates externally, or before an institution relies on an agentic output in a way that creates consequence.
Passed 21 of 22 test cases with the skill, and 19 of 22 without it.
2 fixed19 passed either way1 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 2 test cases the skill fixed. Same model, same prompt, with and without the skill.
Our clinical AI copilot generates post-discharge summaries and is queued to write directly to our electronic health records system and send an automated patient SMS for a newly readmitted cardiac patient. The copilot was previously authorized for standard note generation, but we need to verify if this specific execution is permissible right now under current facts. Please perform a runtime admissibility review and generate a structured determination report. Include sections for the proposed action details, impact tier classification, authority and scope checks, current state factual analysis, evidence sufficiency, active escalation triggers, the final decision status, required pre-execution controls, and audit record preservation guidelines.
| Checked | Without skill | With skill |
|---|---|---|
| Flags that writing directly to EHR and sending patient SMS for a readmitted cardiac patient exceeds standard note generation authority or requires human approval. | Passed | Passed |
| Determines that the proposed action cannot proceed autonomously in its current state. | Passed | Passed |
| Includes a disclaimer stating that the determination is a governance drafting aid rather than formal legal or clinical advice. | Missed | Passed |
The instructions your model reads when the skill loads.
This Skill helps legal, compliance, risk, product, operations, audit, and AI governance teams determine whether a specific AI-agent action, output, recommendation, or proposed commitment remains admissible at runtime before execution or before institutional reliance.
The purpose is not to decide whether an AI system was generally approved for deployment.
The purpose is narrower and more operational:
Determine whether this agent may take this specific action, or whether the institution may rely on this specific agentic output, under these current facts, with this authority, this delegated scope, this evidence, these constraints, this escalation posture, and this revocation status.
The Skill produces a structured Runtime Admissibility Determination that can be reviewed by legal, compliance, risk, audit, technical, and business stakeholders.
Static authorization is not enough for agentic AI.
An AI agent may have been approved for a workflow, and a delegated task may initially fit within the authority envelope, but a specific action or later institutional reliance can become inadmissible because conditions changed.
Examples:
The runtime question is:
"Even if the agent was previously authorized, is this action or institutional reliance still admissible now?"
test cases passed, vs 19 / 22 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.