repo-delivery
Pass
Audited by Gen Agent Trust Hub on Jul 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use standard Git CLI tools (
git status,git commit,git push, etc.) for repository delivery. The execution is bounded by strict 'Hard Rules' that require explicit path staging and verification of the staged diff before mutation. - [PROMPT_INJECTION]: The skill evaluates repository-specific configuration files such as
AGENTS.mdandCLAUDE.md. While these serve as an indirect prompt injection surface for untrusted repositories, the skill mitigates this risk by requiring the agent to prioritize hard-coded safety constraints and user authorization over external guidance. - [SAFE]: No evidence of obfuscation, hardcoded credentials, or unauthorized network operations was found. The skill includes documentation for its own maintenance referencing the author's GitHub repository (
github.com/idaibin/skills), which is consistent with the provided vendor context.
Audit Metadata