source-command-recall-update

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

• Third-party content exposure detected (high risk: 0.75). SKILL.md:16-18 and 67-72 describe fetching GitHub release JSON and using the release body text (outsider-authored public web content) as LLM-readable prose to print an excerpt into the agent context/output at runtime.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill issues a runtime curl to https://api.github.com/repos/edheltzel/Recall/releases/latest to fetch the latest release tag/body which is then parsed and presented to the user, meaning remote content directly controls the agent's output.