The Agent Skills Directory

[SAFE]: The skill demonstrates security-conscious design by requiring a mandatory human approval step (Step 5) before filing any issues to GitHub. This human-in-the-loop mechanism effectively mitigates risks associated with automated content generation.
[SAFE]: Instructions explicitly state that secret values must stay out of the generated specifications, preventing accidental credential exposure.
[SAFE]: The skill uses standard tools (git, gh CLI) for its operations within a clearly defined scope (issue creation and record tracking) and does not attempt to execute remote code or perform unauthorized privilege escalations.

create-spec