walkthrough-script-agent
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data, specifically app codebases and screen recordings, which could contain malicious instructions designed to manipulate the agent's output.
- Ingestion points: App codebase, screenshots, and screen recordings provided by the user (described in
SKILL.md, Step 2). - Boundary markers: The instructions for the
feature-analyzeragent do not define clear boundaries or instructions to ignore embedded commands within the analyzed material. - Capability inventory: The skill can read local files, write new markdown and JSON files, and execute a Python-based frame extraction tool.
- Sanitization: There is no evidence of input sanitization or filtering of the content being analyzed.
- [COMMAND_EXECUTION]: The skill executes a local shell command to process video recordings for scene detection.
- Evidence:
SKILL.mdcontains a command to runextract_frames.pyfrom a related skill directory.
Audit Metadata