gimp
Pass
Audited by Gen Agent Trust Hub on May 5, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using a specific CLI harness (
cli-anything-gimp) to interact with GIMP. These commands include file manipulation, geometry adjustments, and batch processing. - [PROMPT_INJECTION]: Evaluated for indirect prompt injection vulnerability surfaces in processing academic figures. The skill interpolates variable identifiers like
{citationKey}into shell command arguments. - Ingestion points: Reads image files and metadata (citation keys) from potential untrusted sources like academic PDFs.
- Boundary markers: Absent; the instructions do not explicitly mandate delimiters or sanitization for interpolated path variables.
- Capability inventory: Includes file read/write operations and shell command execution via
cli-anything-gimpas described in SKILL.md. - Sanitization: No specific sanitization or validation steps are provided in the instructions for the
{citationKey}variable, relying on the agent's underlying safety filters for shell command execution.
Audit Metadata