install-hooks

Pass

Audited by Gen Agent Trust Hub on Oct 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses shell commands to install and configure the ggshield CLI tool. This includes using platform package managers like brew, choco, apt, and dnf, as well as executing the ggshield install command to wire hooks into git and AI tools.
  • [EXTERNAL_DOWNLOADS]: Fetches the ggshield CLI and its official installation scripts (install.sh, install.ps1) from GitGuardian's GitHub repositories and Cloudsmith. These downloads are from the tool's official vendor and are used to establish the required environment.
  • [PRIVILEGE_ESCALATION]: Employs sudo for administrative tasks such as installing package artifacts (e.g., .pkg, .deb, .rpm). The skill explicitly instructs the agent to obtain informed user consent before executing global (--mode global) installations that modify the user's system-level git configuration.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests user prompts to determine installation parameters. It mitigates potential instruction injection risks by requiring the agent to stop and clarify ambiguous requests and by routing actions through a defined command-line interface rather than directly interpreting untrusted input as shell code.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 23, 2026, 12:42 PM