setup-skills
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill is designed with a security-first approach, specifically implementing 'git guardrails' by modifying
.claude/settings.json. These guardrails prevent the AI agent from performing destructive Git operations (e.g.,git push,git reset --hard,git clean), thereby protecting the repository's integrity and history. - [COMMAND_EXECUTION]: The skill instructs the agent to execute local CLI tools such as
git,gh,glab, andjq. These are used for repository exploration and configuration management (e.g., creating triage labels or merging JSON settings). The execution is localized to the user's project environment and well-known service providers (GitHub/GitLab). - [INDIRECT_PROMPT_INJECTION]: The skill reads project metadata files like
package.jsonand.git/configto determine the repository's state. While these files are technically untrusted external inputs, the skill uses this information for basic discovery logic and does not interpolate the data into executable commands or prompts in a way that would allow for injection attacks.
Audit Metadata