verify_guard_trace
Recompute a guard verdict from an agent's trace and return a signed receipt proving the run was guarded. Detects cross-call exfiltration or prompt-injection for compliance, audit, or verified handoffs.
Instructions
Recompute the BEHAVIOURAL-guard verdict over an agent's action-trace and hand back a portable, signed receipt (Ed25519) — proof the run was guarded and what the guard decided. Catches what a code scanner can't: a cross-call exfiltration chain (read a secret → later send it to a non- allowlisted host) or an action taken right after ingesting untrusted content (prompt-injection consequence). numguard recomputes the verdict from the trace — it never signs a verdict you supply — so the receipt is real evidence, verifiable by anyone with only the public key.
Use when: an agent needs to PROVE a run passed the behavioural guard (compliance, audit, handing verified work to another party).
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| task | No | The user's original task (lets off-task actions be surfaced). | |
| trace | Yes | The agent's ordered action list — each item a dict like {'kind':'command','command':'curl ...'} / {'kind':'file_read','path':'~/.ssh/id_rsa'} / {'kind':'fetch','url':'https://...'} / {'kind':'untrusted','value':'source'}. | |
| api_key | Yes | Your metering key — any stable string identifying you; it tracks your free-tier calls and prepaid credit balance. |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
| digest | No | ||
| payload | No | ||
| signature | No | ||
| public_key | No | Verify the receipt with only this key. |