credential-switch
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes local scripts 'credential-manager.py' and 'execution-tracker.py' via command line to switch and validate brand profiles.
- [PROMPT_INJECTION]: The skill possesses an indirect injection surface by reading brand profile JSON files. Ingestion points: '~/.claude-marketing/brands/{slug}/profile.json'. Boundary markers: None. Capability inventory: subprocess calls to 'credential-manager.py'. Sanitization: The skill verifies the existence of the brand directory on the filesystem before execution.
- [DATA_EXFILTRATION]: It accesses credential files in '~/.claude-marketing/', which is necessary for its purpose as a credential manager. No evidence of remote exfiltration was found.
Audit Metadata