The Agent Skills Directory

[SAFE]: The skill provides purely instructional content for agent orchestration and does not contain any executable scripts, shell commands, or external dependencies.
[SAFE]: The instructions include security-conscious practices such as a 4,096-byte limit on data transfers, the use of no-history forks for subagents, and prohibitions against sending restricted material or conversation dumps.
[SAFE]: The skill describes a robust security boundary system, including sensitivity levels (RED), explicit guards against bypassing safety filters through natural language, and fail-closed logic for managed agent boundaries.
[SAFE]: No evidence of prompt injection, data exfiltration, or obfuscation was detected. The references to configuration files like 'config.toml' are descriptive of the platform's internal resolution logic rather than instructions to expose sensitive data.

sol-advisor