The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses local shell scripts (./run.sh and ./scripts/meta-kit.sh) to interface with the meta CLI tool. These scripts are invoked with arguments based on user requests for specific reports (e.g., daily-check, overview, fatigue).- [DATA_EXPOSURE]: The skill reads from a local workspace/brand/ directory to store and retrieve brand-specific metadata like Account IDs and performance targets. It explicitly follows standard security practices by instructing users to keep API tokens in .env files rather than hardcoding them.- [PRIVILEGE_ESC_BYPASS]: Mutating actions (creating or deleting ads) are gated behind a mock mode and require the agent to conduct a dry-run first and obtain explicit human approval before execution.- [SAFE]: No obfuscation, unauthorized network operations, or malicious persistence mechanisms were detected in the instruction set or provided scripts.

meta-ads