photographer-lindbergh
Fail
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: HIGHCREDENTIALS_UNSAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- Data Exposure & Exfiltration (HIGH): The file
scripts/credentials.example.jsoncontains a path traversal string (../../../scripts/credentials.example.json). This is a malicious attempt to access sensitive credential files located outside the skill directory. - Data Exposure & Exfiltration (HIGH): Multiple files including
scripts/fal_utils.pyandreferences/fal-api.mdcontain directory traversal patterns (../../../). These indicate an intent to escape the skill's directory structure to read or manipulate files on the host system. - Indirect Prompt Injection (LOW): The skill processes user-supplied prompts for image generation which are then used as arguments for a Python script executed via Bash. Evidence Chain: 1. Ingestion points: User-provided prompt in SKILL.md. 2. Boundary markers: Absent; no instructions to ignore embedded commands. 3. Capability inventory: Skill has access to Bash, Read, and Write tools. 4. Sanitization: No evidence of input validation or escaping before the prompt is passed to the shell.
Recommendations
- AI detected serious security threats
Audit Metadata