setup-custom-skills
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
git remote -vcommand to discover the repository's remote host. This is a standard discovery operation used to recommend the appropriate issue tracker configuration (GitHub, GitLab, or local). - [SAFE]: The skill performs file system operations to read and update repository configuration files like
CLAUDE.md,AGENTS.md, and project-specific documentation within thedocumentation/agents/directory. These actions are transparent and aligned with the primary purpose of environment setup. - [SAFE]: The generated configuration templates for GitHub and GitLab utilize official CLI tools (
ghandglab) to perform repository management tasks. These are well-known, legitimate developer tools. - [INDIRECT_PROMPT_INJECTION]: The skill sets up a workflow where the agent will eventually ingest data from issue trackers and pull requests. While this creates an attack surface for indirect prompt injection in the resulting workflow, the setup process itself does not execute untrusted data, and the configuration focuses on enabling standard operational visibility.
Audit Metadata