The Agent Skills Directory

[SAFE]: The skill provides technical documentation and code for the CrewAI framework. All mentioned packages (crewai, pydantic) and domains (crewai.com) are official and trusted resources. Mentions of command execution (crewai install, crewai run) refer to the framework's standard CLI tools.
[INDIRECT_PROMPT_INJECTION]: The skill describes how to build flows that ingest data from external sources, which represents a potential attack surface for indirect prompt injection.
Ingestion points: Data enters the flow through 'kickoff(inputs=...)', 'handle_turn(message=...)', and '@human_feedback' mechanisms documented in SKILL.md and its references.
Boundary markers: The provided code examples do not specify the use of delimiters or boundary markers for these inputs.
Capability inventory: Flows are designed to invoke 'Crew' and 'Agent' instances, which are capable of executing tasks and potentially using tools for network or file system operations.
Sanitization: No specific input validation or sanitization logic is presented in the examples.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill promotes security best practices for credential management, specifically advising users to store API keys in environment variables or '.env' files rather than hardcoding them.

build-flow