rmbc-copy-audit
Pass
Audited by Gen Agent Trust Hub on Jun 29, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted user data (
copy_text) intended for auditing without implementing safety boundaries or delimiters. - Ingestion points: External content is ingested through the
copy_textinput parameter defined in bothSKILL.mdandSKILL-standard-archived.md. - Boundary markers: Absent. The instructions do not employ XML tags, triple quotes, or specific delimiting tokens to help the model distinguish between the audit instructions and the text being audited.
- Capability inventory: The skill is restricted to text generation and does not list any allowed-tools that would permit file system manipulation, network exfiltration, or command execution.
- Sanitization: No sanitization or validation mechanisms are described to mitigate the risk of instructions embedded within the copy influencing the agent's behavior.
Audit Metadata