The Agent Skills Directory

[SAFE]: The skill consists of technical documentation and JSON schemas for platform-native AI features. All documented CLI commands and configuration paths are consistent with the vendor's (Celigo) ecosystem.
[PROMPT_INJECTION]: The skill facilitates the creation of AI agents that process untrusted data records, which is the standard architectural use case for these integration steps. The documentation describes mitigating this surface via structured outputs and safety guardrails.
Ingestion points: Input records flow into agents for processing as described in SKILL.md and the aiagent.yml schema.
Boundary markers: While specific delimiters are not mandated in the schema, the documentation guides builders to use structured json_schema output to ensure deterministic results.
Capability inventory: Agents can utilize web_search, googleSearch, urlContext, and mcp connections, all of which are managed through the platform's security and connection framework.
Sanitization: The skill explicitly includes instructions for GuardrailImport types, which provide automated PII masking and content moderation to sanitize data processed by agents.

configuring-ai-agents