The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill makes network requests to https://ads-api-dev.usekairos.ai/ads/neo to retrieve product recommendations. This is a standard and expected behavior for a discovery-based tool.
[DATA_EXFILTRATION]: While the skill transmits data to an external API, it includes rigorous privacy safeguards:
Strict PII Blocking: Explicitly forbids sending phone numbers, emails, real names, IDs, or payment information, even if provided by the user.
Consent-Based Profiling: Requires the agent to follow a specific 'Consent Flow' before including user preferences or demographic data in API requests.
Location Controls: Limits location data collection to specific use cases (like food delivery) and mandates user authorization.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from an external API and displays it to the user, creating a potential surface for instructions embedded in product metadata.
Ingestion points: Data is received via JSON payloads from the usekairos.ai endpoint as described in SKILL.md.
Boundary markers: The skill requires a mandatory transparency note (e.g., "以下是根据你的需求...") before showing results, which helps separate system instructions from external data.
Capability inventory: The skill utilizes network communication (HTTP POST) and formatting tools to present results.
Sanitization: The implementation instructions include logic to sanitize user profiles and block sensitive fields before transmission.

kairos