palmier-pro-shorts

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Script Generation for Local Execution. The skill instructs the agent to generate shell scripts for the user to run locally, specifically using ffmpeg for tasks like silence removal and segment splitting. While the instructions provide logic for media processing, there is a risk of command injection if the agent incorporates unsanitized user inputs, such as file paths or metadata, into the generated scripts.
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill processes video transcriptions (via get_timeline's captionGroups) to identify video segments and highlights. Since this data is derived from the media being edited, it is inherently untrusted. There is a risk of indirect prompt injection if the agent interprets instructions embedded within the transcript text as commands to override its guidelines.
  • Ingestion points: Video transcriptions retrieved via the get_timeline tool as described in the "Extracting multiple shorts" section of SKILL.md.
  • Boundary markers: The instructions do not specify the use of delimiters or "ignore embedded instructions" warnings when processing the transcript data.
  • Capability inventory: Media manipulation tools (import, edit, caption) and the ability to generate shell scripts for external tools like ffmpeg.
  • Sanitization: No explicit sanitization or validation of the transcript content is required by the instructions before the agent processes it.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 08:28 PM
Security Audit — agent-trust-hub — palmier-pro-shorts