The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill has ingestion points for audience data and performance evidence as specified in references/skill-spec.json. It mitigates potential risks through a process-based 'Evidence discipline' that mandates the classification of all material claims (fact, inference, assumption) and strictly prohibits the fabrication of research or tool evidence.
[COMMAND_EXECUTION]: The skill references local routing logic and specialist agents (e.g., dynamic_router.py, creative-strategist.md) via relative filesystem paths. These components are used for workflow orchestration and specialist selection within the intended operating environment and do not involve unauthorized command execution.

content-strategy