transition-handoff

Pass

Audited by Gen Agent Trust Hub on Oct 19, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input (the "mind dump") to generate narratives and video scripts. This creates a surface where instructions embedded in the user's input could attempt to influence the agent's behavior.
  • Ingestion points: User-provided "mind dump" text and dictation in SKILL.md.
  • Boundary markers: The instructions lack explicit delimiters to separate user input from the agent's internal logic.
  • Capability inventory: Includes reading local files (paths, line numbers), viewing command history (last error, last command), and executing video rendering commands.
  • Sanitization: No specific filtering or escaping is applied to the dictated content before it is used in the narrative or script.
  • [EXTERNAL_DOWNLOADS]: The skill directs the agent to install the making-videos dependency from the author's GitHub repository using npx skills add JustYannicc/skills --skill making-videos if it is not already present.
  • [COMMAND_EXECUTION]: The instructions include executing commands to render video files to the local file system (~/Movies/Handoffs/) and using npx for dependency management. The agent is also instructed to access workspace metadata such as current Git branch, file paths, and recent command errors to enrich the handoff content.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 19, 2026, 06:48 PM