service
Fail
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: HIGHCREDENTIALS_UNSAFEDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
- CREDENTIALS_UNSAFE (HIGH): The helper script
scripts/railway-api.shreads sensitive credentials (user.token) from the Railway configuration file located at~/.railway/config.json. - DATA_EXFILTRATION (HIGH): The authentication token is sent to the external domain
backboard.railway.comviacurl. Although this is the official Railway API, the pattern of reading and transmitting secrets is classified as high severity. - COMMAND_EXECUTION (LOW): The skill frequently executes bash scripts and the
railwayCLI to perform its functions. - PROMPT_INJECTION (LOW): The skill is vulnerable to Indirect Prompt Injection (Category 8). Evidence: 1. Ingestion points:
railway status --jsonoutput. 2. Boundary markers: No delimiters or warnings used for external data. 3. Capability inventory: API interaction viascripts/railway-api.shallowing service creation and modification. 4. Sanitization: None present for the content of ingested data fields.
Recommendations
- AI detected serious security threats
Audit Metadata