larrybrain

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

• Third-party content exposure detected (high risk: 0.85). The required workflow downloads third-party skill files (including SKILL.md free text) from the LarryBrain API at runtime and then the agent reads and follows those instructions, so outsider-authored markdown content is ingested into the LLM context.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill instructs the agent to download and run skill files at runtime using the install API (e.g., https://www.larrybrain.com/api/skills/install?slug=SLUG&mode=files), and those fetched files (SKILL.md + supporting files) directly control agent behavior and must be followed, so this is a runtime external dependency that controls prompts/instructions.