quest
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill accepts arbitrary user input via the
$ARGUMENTSparameter and uses it to drive a recursive loop of subagent spawning. There are no built-in constraints or safety checks on the content of these objectives, allowing potentially malicious instructions to be executed with high persistence. - [INDIRECT_PROMPT_INJECTION]: The skill acts as a passthrough for untrusted data into an autonomous process.
- Ingestion points: The
$ARGUMENTSfield inSKILL.mdis the primary ingestion point for external data. - Boundary markers: The skill lacks delimiters (e.g., XML tags or clear separators) to prevent the subagent from interpreting parts of the objective as system-level commands.
- Capability inventory: The skill is capable of spawning subagents and verifying their output repeatedly.
- Sanitization: No sanitization, escaping, or validation logic is present to inspect the objective before execution.
Audit Metadata