reference-decode
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data sources to extract prompt structures.
- Ingestion points: Processes benchmark videos, contact sheets, frame sets, and rough ideas as described in the Use When and Default Workflow sections of SKILL.md.
- Boundary markers (absent): The instructions lack explicit markers to isolate the untrusted source content from the agent's core instructions.
- Capability inventory: The skill utilizes the 'postplus' CLI for task execution and confirmation logic.
- Sanitization (absent): No sanitization or validation logic is defined for the content extracted from the external reference materials.
- [COMMAND_EXECUTION]: The skill contains logic to interact with the host system via the 'postplus' command-line utility.
- Evidence: Specifically references the command 'postplus quote confirm --json --challenge-file' to resolve quote-confirmation challenges during execution.
Audit Metadata