The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional text guiding an AI agent through a meta-workflow step (Step ②). It establishes a process for human-in-the-loop verification of project goals.
[NO_CODE]: The skill does not include or invoke any scripts, binaries, or shell commands. It references other internal workflow steps and a local document ('references/quality-bar.md'), which is standard practice for organizational knowledge management.
[SAFE]: No obfuscation, prompt injection patterns, or data exfiltration mechanisms were detected in the instructions. The 'anti-lowering rule' and human sign-off requirement act as process-level safeguards against quality degradation.

freeze-acceptance