openobserve

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt contains plaintext example credentials (e.g., curl -u 'root@example.com:Complexpass#123', Docker env vars with ZO_ROOT_USER_PASSWORD, Fluent Bit/Prometheus password fields) and shows them being embedded directly in commands/configs, which forces or encourages the LLM to include secret values verbatim in generated output.

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

• Third-party content exposure detected (high risk: 0.95). The required runtime workflow includes WebFetch("https://openobserve.ai/docs/quickstart/") (and additional doc page fetches), which pulls public web content authored by outsiders into the agent’s LLM context via the fetched page text.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly requires a runtime fetch via WebFetch("https://openobserve.ai/docs/quickstart/") (and other pages under https://openobserve.ai/docs/) before performing any work, meaning external content is fetched at runtime and directly controls the agent's instructions.