gifgrep
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill installs the 'gifgrep' binary from a third-party GitHub repository (steipete/gifgrep) using either Homebrew or Go.
- [COMMAND_EXECUTION]: The skill requires the execution of the 'gifgrep' CLI tool to perform network searches, file downloads to the user's local directory, and image processing.
- [CREDENTIALS_UNSAFE]: The skill documentation identifies the use of 'GIPHY_API_KEY' and 'TENOR_API_KEY' environment variables for authentication, which the agent must access to function.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted metadata from external providers. 1. Ingestion points: Search results from Tenor and Giphy (e.g., titles, tags, and JSON payloads). 2. Boundary markers: Absent; external data is processed directly without explicit delimiters. 3. Capability inventory: Includes file system write access to '~/Downloads' and shell execution of the 'gifgrep' binary. 4. Sanitization: No sanitization or validation of the external metadata is described in the skill.
Audit Metadata