uithub-fetcher
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes the
uithubCLI using parameters like repository names and search patterns derived from user-provided URLs. This presents a potential command injection risk if input is not properly handled by the agent context. - [PROMPT_INJECTION]: The skill allows for indirect prompt injection through the ingestion of external repository data.
- Ingestion points: Repository code, issues, and pull request data are retrieved via the
uithubcommand as described inSKILL.md. - Boundary markers: There are no defined delimiters to separate external content from internal instructions.
- Capability inventory: The agent can execute CLI tools and perform repository searches.
- Sanitization: No content sanitization or filtering is implemented for the fetched GitHub data.
- [EXTERNAL_DOWNLOADS]: The skill requires the
uithubCLI and directs users to install it via a local path reference. - [CREDENTIALS_UNSAFE]: The tool stores sensitive GitHub OAuth tokens in a local file at
~/.uithub/token.json.
Audit Metadata