The Agent Skills Directory

[SAFE]: The skill operates as a strategic advisor within a modular AI agent framework. Its instructions focus on applying marketing frameworks (RFM, LTV, JTBD) to provided brand data.
[DATA_EXPOSURE]: The skill contains explicit safety guardrails in its Context Matrix, stating it must 'Never' access customer PII or unmasked purchase records. It uses a structured 'Brand Brain' approach for accessing non-sensitive product and audience data.
[COMMAND_EXECUTION]: There are no shell commands, subprocess calls, or privilege escalation attempts. The skill's operations are limited to data analysis and report generation based on markdown templates.
[EXTERNAL_DOWNLOADS]: No external packages or scripts are downloaded or installed. All logic is contained within the provided markdown references and platform-level system protocols.
[INDIRECT_PROMPT_INJECTION]: While the skill ingests external brand data (products.md, audience.md), it does so within a controlled 'Supervisor' context and uses specific boundary markers defined in its output protocol to prevent instructions in data from influencing agent behavior. It lacks the network or system-level capabilities required to weaponize such an injection.

afa-retain