git-pushing-fast-cn
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses standard Git CLI commands (
git status,git add,git commit,git push) to perform its primary function. These are expected for a Git automation tool. - [DATA_EXFILTRATION]: Includes proactive safety measures to prevent accidental data exposure. Specifically, it instructs the agent to ensure that sensitive files—such as
.env, credential files, private keys, or tokens—are not staged or committed. - [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface as it analyzes repository diffs to generate commit messages. While a malicious diff could attempt to influence the commit message content, the instructions limit the agent's output to Conventional Commit formats and require logical分段 (sections), reducing the impact of such attempts. This is considered a standard operational risk for this type of tool.
- [SAFE_PRACTICES]: The skill enforces several best practices: it forbids the use of
--no-verifyto bypass hooks, prevents modifications to global Git configurations, and prohibits destructive commands likegit reset --hardorgit checkout -- <file>unless explicitly requested by the user.
Audit Metadata