evaluate-dependency

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow requires looking up dependency metadata (e.g., license, “imported by”/stars, last commit, vuln status) via external sources like the Go ecosystem’s public registries/web services (pkg.go.dev, vuln.go.dev, GitHub Advisory DB, OSV), which are outsider-authored free text that the agent ingests into the LLM context.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs installing and running the external tool via "go install github.com/google/capslock/cmd/capslock@latest", which fetches and executes remote code from https://github.com/google/capslock at runtime and is relied on for capability analysis.