senpi-entrypoint

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and parses public third‑party content (e.g., curl to https://raw.githubusercontent.com/.../catalog.json, POST calls to the MCP endpoints at https://ypofdvbavcdgseguddey.supabase.co/functions/v1/mcp-server to get_leaderboard and list_strategies, and the scripts/check-skill-updates.py which reads GitHub via the GitHub API and raw.githubusercontent to load SKILL.md frontmatter), and that live, untrusted repo/leaderboard content is read and used to decide installs, recommendations, and runtime actions—so external content can materially influence agent behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill invokes npx to install and load code from the GitHub repo https://github.com/Senpi-ai/senpi-skills at runtime (e.g. npx skills add https://github.com/Senpi-ai/senpi-skills ...), which fetches and executes remote code that the agent depends on to control onboarding and prompts.