add

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (medium risk: 0.30). The skill instructs the agent to pipe remote scripts from an arbitrary third-party site into bash as part of its fallback mechanism.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (critical risk: 1.00). The skill fetches and executes remote scripts from a third-party server via bash execution, representing a remote code execution vulnerability pattern characteristic of malicious skills.

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

• Third-party content exposure detected (low risk: 0.10). The workflow fetches capability catalogs and documentation from a remote Convex site (https://basic-anteater-667.convex.site), which constitutes fetching remote content/documentation rather than ingesting monitored outsider-authored free text.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill fetches executable bash scripts (add-hosting, add-component) and instruction documents from an arbitrary third-party site (basic-anteater-667.convex.site) to govern its runtime behavior and execution, representing a weak, unverifiable operational dependency.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill instructs the agent to download and execute remote scripts via curl piped to bash from an external URL, which compromises the security state of the machine.