The Agent Skills Directory

[SAFE]: The skill operates as a structured prompt engineering framework. It coordinates several sub-agents by passing text-based instructions and context retrieved from local workspace files (e.g., .agents/mkt/imc-plan.md). No executable code, scripts, or binary files are included.- [SAFE]: Data handling is restricted to the agent's local environment. The instructions specifically guide the agent to read internal marketing artifacts to ground the generated content, with no evidence of attempts to access sensitive system files (like SSH keys or environment variables) or exfiltrate data to external servers.- [SAFE]: The 'Dispatch Protocol' described in the orchestrator instructions is a set of guidelines for how the LLM should format prompts for its sub-agents. It does not involve dynamic code evaluation (eval/exec) or shell command injection.- [SAFE]: The use of a 'Critic Agent' as a quality gate is a security best practice for autonomous agents, ensuring that the final output adheres to predefined constraints and character limits, which helps prevent accidental or malicious prompt overflows in downstream tasks.

content-create