caveman-compress
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted content from user-specified files for the purpose of compression, creating a surface for injection attacks.
- Ingestion points: Content from the file at the path provided in the
<absolute_filepath>argument is read into the agent's context. - Boundary markers: Absent. The instructions do not provide delimiters or specific commands to ensure the LLM ignores instructions that might be embedded within the files being compressed.
- Capability inventory: The skill is capable of executing local Python scripts (
python3 -m scripts), reading files, creating backups, and overwriting existing files. - Sanitization: Absent. The skill does not describe any validation or sanitization of the input file content before it is processed by the LLM.
- [COMMAND_EXECUTION]: The skill triggers the execution of a local Python module which performs the compression logic.
- Evidence: The skill instructions explicitly direct the agent to run
python3 -m scripts <absolute_filepath>.
Audit Metadata