The Agent Skills Directory

[SAFE]: The skill uses a rigorous multi-agent orchestration framework (Strategist, Composer, Format Checker, Voice Auditor, Critic) to manage the generation and validation of ad copy. This structure provides internal oversight and ensures that content is checked for policy compliance and formatting before delivery.
[SAFE]: Strict enforcement of Meta's Advertising Standards is integrated into the workflow via agents that verify character limits, substantiation of claims, and the absence of banned wording (policy-floor.md).
[PROMPT_INJECTION]: The skill ingests data from local research files such as product-context.md and icp-research.md. While this constitutes an indirect prompt injection surface, the risk is negligible as the agents are instructed to treat this data as passive context for copywriting rather than a source of commands. The presence of validation gates further mitigates potential risks from untrusted input.
[COMMAND_EXECUTION]: Although the Bash tool is enabled, it is used only for legitimate file management operations such as writing markdown artifacts to the local directory, consistent with the skill's stated purpose.
[EXTERNAL_DOWNLOADS]: External links found in the reference material point exclusively to official Meta documentation and legitimate practitioner insights on established platforms (x.com), serving purely as documentation sources.

ad-copy