The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is configured to invoke external Python scripts (../../scripts/dynamic_router.py, ../../scripts/neural_router.py, ../../scripts/skill_router.py) found in the framework's parent directory to manage complex task dependencies and specialist selection.
[EXTERNAL_DOWNLOADS]: The skill specification includes the capability to perform web research for real-time information when required for the decision-making process.
[INDIRECT_PROMPT_INJECTION]: The skill processes behavioral data which could potentially contain malicious instructions. However, it incorporates strong mitigation patterns such as mandatory evidence classification (fact vs. inference), boundary framing, and the requirement to consider credible alternatives before finalizing recommendations.

behavioral-marketing