sillytavern-honcho-setup

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly asks the user to paste their Honcho API key and workspace ID and instructs the agent to "surface both values" (e.g., embedding $HONCHO_API_KEY/$HONCHO_WORKSPACE_ID in browser instructions), which requires the LLM to handle and output secret values verbatim.

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

• Potentially malicious external URL detected (high risk: 0.85). This skill installs and executes the Honcho SDK (npm install of @honcho-ai/sdk) and then performs real runtime requests that rely on the Honcho service (see https://app.honcho.dev), so a remote service/package fetched at install/runtime can supply data that directly influences prompts/agent behavior and executes code inside the plugin.