postman
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHCREDENTIALS_UNSAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- Indirect Prompt Injection (HIGH): The skill processes untrusted external data from API responses. 1. Ingestion points:
pm.response.json()in SKILL.md. 2. Boundary markers: Absent. 3. Capability inventory:pm.sendRequest(network),npx newman(shell execution),pm.environment.set(state modification). 4. Sanitization: Absent. - Data Exposure & Exfiltration (HIGH): The skill explicitly handles
apiKeyandclientSecret. It provides code to send these viapm.sendRequestto a dynamictokenUrl, creating a vector for credential exfiltration if an attacker influences the environment. - Command Execution (MEDIUM): The use of
npx newmaninvolves executing a CLI tool that may download packages at runtime. This introduces a risk of remote code execution if the package source is compromised or if typosquatting occurs.
Recommendations
- AI detected serious security threats
Audit Metadata