opencode

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill prompt contains multiple code/config examples that embed API keys or secret-looking values directly (env files, Authorization headers, apiKey fields, server key arrays), which would require the LLM to handle or emit secret values verbatim.

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

• Third-party content exposure detected (medium risk: 0.30). SKILL.md describes OpenCode skill/plugin event handling where plugin code can process runtime SSE message.part text (from user-authored session messages) and even parse tool-call JSON embedded in that text via checkForToolCallAsText(part.text)/handleEvent.