theangrypit-skillset-bootstrap

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

• Third-party content exposure detected (high risk: 0.85). At runtime the script executes npx skills add <source> ... for multiple outsider-authored public GitHub repos (e.g., https://github.com/TheAngryPit/TheAngrySkills.git, https://github.com/vectorize-io/hindsight.git), and the npx skills installer will fetch and ingest their README/skill text into the agent’s LLM context, enabling indirect prompt injection via those repo-authored free text.

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

• Potentially malicious external URL detected (high risk: 1.00). The install/update scripts invoke npx skills add on external GitHub repositories (e.g. https://github.com/TheAngryPit/TheAngrySkills.git, https://github.com/vectorize-io/hindsight.git, https://github.com/mattpocock/skills.git), which at runtime fetches remote skill code/content that can contain prompt templates or executable skill code that will control agent behavior.