peon-ping-remix
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it reads user-controlled 'caption' data from an external job JSON file and incorporates it into new render prompts. These prompts are subsequently saved to the filesystem and processed by a rendering script. Maliciously crafted captions could attempt to override agent instructions or influence the content of generated files. • Ingestion points: Job JSON file read from a path provided during invocation (specifically the 'caption' field). • Boundary markers: None identified; the skill does not use delimiters to isolate the untrusted caption data. • Capability inventory: The skill has the ability to write JSON files to the local disk and execute shell commands via python3. • Sanitization: No explicit sanitization or validation of the caption content is performed before interpolation into new prompts.
- [COMMAND_EXECUTION]: The skill performs shell command execution by running 'python3' to invoke a local rendering script ('pack-render.py'). The path to this script is resolved dynamically at runtime using environment variables and local configuration paths. While this is the intended core functionality of the skill, subprocess execution with dynamically resolved paths is a sensitive operation.
Audit Metadata