rdd-research

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow (Step 2: "Research — via web search" in SKILL.md and multiple "Method: Web research" entries in docs/research-log.md) instructs the agent to fetch and interpret open web / community sources and to use those findings to drive routing, evaluation, and promotion decisions, so untrusted third-party content can directly influence tool use and next actions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill includes a runtime git clone of a remote repository (https://github.com/mrilikecoding/llm-orchestra-library.git) which, when fetched, contains ensemble YAMLs that directly define agent prompts/behavior and thus can control agent instructions at runtime.