debuzz
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute the
agyCLI tool via a shell command. It specifically directs the agent to use shell command substitution ($(cat ...)) to pass the contents of a temporary file as an argument to the tool. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content (either from the agent's previous responses or user-pasted text) and incorporates it into a prompt for an external AI translation tool.
- Ingestion points: Untrusted data is retrieved from the conversation history or direct user input as described in
SKILL.md. - Boundary markers: The instructions do not specify any delimiters or safety warnings to help the external tool distinguish between the text to be translated and potential instructions embedded within that text.
- Capability inventory: The skill uses the
Writetool to create input files and theshellcapability to execute commands. - Sanitization: There is no evidence of input validation, escaping, or filtering of the source text before it is passed to the command-line tool.
Audit Metadata