use-railway

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill instructs fetching and relaying credentials (bucket S3 creds, device-code sign-in URL/code) and shows setting variables via direct CLI args (KEY=value), which requires the agent to output secret/auth token values verbatim and thus risks secret exfiltration.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.65). SKILL.md routes “Analyze a database” to the database analysis scripts, which (per references/analyze-db.md) collect and then quote raw recent_logs / slow query text from the target database/SSH/API into the agent’s response/LLM context; those log lines are outsider-authored free text (originating from the user’s deployed service/database, not authored by the operating user).

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill contains runtime installation commands that fetch and execute remote shell scripts (curl -fsSL agents.railway.com | sh and bash <(curl -fsSL https://railway.com/install.sh)), which cause arbitrary remote code to be run during skill setup.