The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources (Korean government safety databases via a proxy), which is a standard surface for indirect prompt injection.
Ingestion points: External data is fetched from various endpoints under k-skill-proxy.nomadamas.org/v1/mfds/ in the SKILL.md workflow.
Boundary markers: The skill includes explicit instructions to perform a mandatory interview first, check for 'red flag' symptoms, and prioritize emergency services over API results. It also mandates that the agent must not provide final medical diagnoses.
Capability inventory: The skill's capabilities are limited to network read operations (GET requests); it does not possess file-write, command execution, or other dangerous system capabilities.
Sanitization: No specific programmatic sanitization of the external API response is defined, though the response policy limits the agent's output to official summaries.
[DATA_EXFILTRATION]: The skill makes network requests to an external proxy to retrieve safety information.
Network operations: Uses curl examples and instructions to query k-skill-proxy.nomadamas.org with user search parameters. This is the core functionality of the skill and does not involve the exfiltration of sensitive local data or credentials.

legal-mfds-safety