upgrade

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

• Third-party content exposure detected (high risk: 0.70). The skill explicitly fetches upstream content from public GitHub (e.g., curl to https://raw.githubusercontent.com/.../package.json and git clone https://github.com/heyitsnoah/claudesidian.git), then diffs and uses an AI-merge workflow that reads and acts on that third-party content as part of its upgrade decisions, exposing the agent to untrusted external content.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill performs runtime fetches from https://github.com/heyitsnoah/claudesidian.git (via git clone) and https://raw.githubusercontent.com/heyitsnoah/claudesidian/main/package.json (via curl) and then uses files from the cloned upstream (including SKILL.md and scripts) to overwrite local skills and hooks, which can directly change agent prompts/instructions and introduce remote code into the user's vault.