extract-source-sample
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill requires the agent to execute shell commands, specifically "git lfs fetch", "git lfs checkout", and the "file" utility. These commands are invoked based on the discovery of files and LFS pointers within the "run-dir" provided by the user, which allows external file content to influence command execution.
- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by processing content from external, potentially untrusted files found in the user-provided directory.
- Ingestion points: Data is ingested from "working/script.json", "HOW_TO_MAKE_THIS_VIDEO.md", and various Python scripts ("working/*.py") located in the user-provided directory.
- Boundary markers: There are no defined delimiters or instructions to help the agent distinguish between data and potential malicious instructions within the ingested files.
- Capability inventory: The agent has the capability to read and write files on the local filesystem and to execute shell commands.
- Sanitization: The skill does not implement validation or sanitization of the extracted content before it is used to update the central character library index or the output JSON metadata.
Audit Metadata