design-systems

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

• Third-party content exposure detected (high risk: 0.75). The workflow can fetch design.md (outsider-authored public web/API content) from https://api.agent-ds.oday-bakkour.com/v1/systems/<slug>/design.md and then uses the resulting free-text file content to instruct the agent, so the LLM may ingest that retrieved prose at runtime.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly fetches remote design files (e.g. curl -fsSL https://api.agent-ds.oday-bakkour.com/v1/systems//design.md) at runtime and then instructs the agent to "Follow it" (including the file's "Agent Prompt Guide"), so fetched content directly controls agent prompts.