find-credential-usage
Fail
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: HIGHCREDENTIALS_UNSAFECOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [CREDENTIALS_UNSAFE]: The skill accesses highly sensitive system files and locations containing credentials.
- Evidence: Searches shell profiles (~/.zshrc, ~/.bashrc, ~/.bash_profile, ~/.profile, ~/.zprofile), environment variables, shell history, and .env files.
- [CREDENTIALS_UNSAFE]: The skill attempts to programmatically extract secrets from the macOS Keychain storage.
- Evidence: Executes the security find-generic-password command to query the system keychain for passwords using the provided search term.
- [COMMAND_EXECUTION]: Shell commands are constructed using raw string interpolation of user-provided parameters, which can lead to command injection.
- Evidence: The variable {{SEARCH_TERM}} is placed directly into multiple bash commands (e.g., grep and security). Malicious input containing shell metacharacters like ; or $(...) could trigger arbitrary command execution.
- [DATA_EXFILTRATION]: Aggregates sensitive credentials from across the system into a single output report.
- Evidence: The procedure instructions require the agent to compile all discovered credentials into a markdown report, concentrating high-value secrets and increasing the risk of accidental exposure or exfiltration via the agent's communication channel.
Recommendations
- AI detected serious security threats
Audit Metadata