k-skill-setup

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt instructs the agent to obtain user API credentials (asking the user or reading secrets.env/vault) and explicitly says the helper will read and send those values verbatim in requests (e.g., ServiceKey query parameter), which requires the LLM to handle/output secret values directly.

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

• Third-party content exposure detected (high risk: 0.70). The SKILL.md explicitly directs the skill to call a default hosted third-party proxy (k-skill-proxy.nomadamas.org) for many public data lookups (e.g., /v1/mfds/food-safety/search, /v1/neis/school-meal, /v1/household-waste/info, KIPRIS, subway/weather endpoints), so the agent will fetch and act on untrusted public third-party content as part of its workflow.