broll-shot-export

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill requires the agent to execute shell-based media commands, specifically calling FFprobe and performing full video decodes. These operations are directed to run through a local executor reference at ../../references/safe-execution.md.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface due to its handling of untrusted external data.
  • Ingestion points: Untrusted data enters the agent context through 'Director shot plans', 'explicit user requests', and 'destination directory' parameters mentioned in SKILL.md.
  • Boundary markers: No delimiters or specific instructions (e.g., 'ignore embedded instructions') are defined to isolate untrusted input from the agent's core instructions.
  • Capability inventory: The skill allows file writes to the broll-production/06-shot-export/ directory and subprocess execution via a platform-provided wrapper.
  • Sanitization: There is no evidence of input validation or sanitization for user-provided timing windows or path strings before they are incorporated into command parameters.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 07:43 AM
Security Audit — agent-trust-hub — broll-shot-export