The Agent Skills Directory

[SAFE]: The skill is primarily educational and referential, providing 31 rules for managing persistent objectives in an AI agent. It does not contain any executable code, malicious shell scripts, or instructions that bypass safety protocols.
[SAFE]: The documentation references installation commands for the '@openai/codex' CLI tool. As OpenAI is a trusted vendor, these references are considered safe and provide legitimate setup instructions for the described technology.
[SAFE]: The skill promotes secure and reliable agent behavior by instructing the model to define explicit boundaries, verify results against concrete evidence, and report blockers honestly, rather than relying on internal model belief.

codex-goals