skills/arcblock/agent-skills/commit/Gen Agent Trust Hub

commit

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [Indirect Prompt Injection] (SAFE): The skill processes untrusted data from the local repository (git diff and status). Evidence: 1. Ingestion points: 'git status', 'git diff', and project history analysis. 2. Boundary markers: Absent. 3. Capability inventory: 'Bash' tool usage for git commands. 4. Sanitization: Not explicitly defined. The risk is mitigated and categorized as SAFE because every state-changing action is gated by a mandatory 'AskUserQuestion' prompt, ensuring the user reviews the proposed changes and commit message.
  • [SAFE] (SAFE): No malicious patterns such as credential exfiltration, remote code execution, or prompt injection were detected. All tools are used for their intended purpose within a local development context.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:34 PM