security-django

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • COMMAND_EXECUTION (SAFE): The skill includes a bash script (scripts/scan.sh) that uses ripgrep (rg) and find to scan local project files for security patterns. All commands are restricted to local file system analysis and do not execute downloaded or untrusted content.
  • DATA_EXFILTRATION (SAFE): The script only performs local read operations for auditing purposes. There are no network calls or mechanisms to send the audited data to external servers.
  • CREDENTIALS_UNSAFE (SAFE): While the documentation discusses SECRET_KEY, it correctly identifies hardcoded keys as a risk and provides remediation guidance to use environment variables instead. No actual credentials are leaked by the skill itself.
  • INDIRECT_PROMPT_INJECTION (LOW): The skill processes user-provided code (Django settings and views). While an attacker could theoretically embed malicious instructions in a code comment to influence the auditor, the skill's narrow focus on specific regex patterns and structured checklists provides a high level of resilience against such attacks.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:18 PM