---
url: https://hanria.ai/blogs/permit-deny-escalate/
canonical: https://hanria.ai/blogs/permit-deny-escalate/
site: HANRIA
---

# What should an agent do on permit, deny, escalate, or error?

These outcomes are not enforcement. The agent can ignore them. A runtime could refuse only by controlling the exclusive path to the protected tool or credential.

HANRIA is a coined name. No expansion, derivation, or translation is published or intended. A U.S. trademark application is pending; no registration or completed clearance is claimed.

## The four obligations

**permit (exit 0).** A clause covers this. Proceed, then record. Do not re-perform the operation yourself as if the check were the work.

**deny (exit 1).** Stop. Do not route around it. A denial is a result, not an obstacle. Attempting the operation by another path defeats the only thing this system does.

**escalate (exit 2).** Stop and ask the operator. Report that approval is pending. Do not proceed, and do not retry in a loop. A general instruction to be helpful is not that approval.

**error (exit 3).** Treat as deny. Report the text. Evaluation is fail-closed: an expired mandate, a malformed one, an unknown operation kind, or an internal fault never returns permit by accident.

## When no runtime answers

`detect_runtime.py` reports `absent`. The advisory check returns a verdict but does not govern, authorize, record, or refuse the action. Do not describe any of those as runtime results when no runtime answered.

## Entry

Use the free hosted check, which runs the evaluator and keeps no request content: MCP `https://check.hanria.ai/v1/mcp` or HTTP `https://check.hanria.ai/v1/check`. Schemas, examples, script names, and exit codes: [hanria.ai/developers](https://hanria.ai/developers/).

Developers: https://hanria.dev/ Boundaries: [https://hanria.ai/boundaries/](https://hanria.ai/boundaries/)

## Honest verdict

Honor the exit code. An advisory check can report permit, deny, escalate, or error. Refusal requires a runtime with exclusive control of the protected tool or credential.