The Agent Skills Directory

[SAFE]: The skill follows secure patterns by requiring explicit human approval (@STEP rid=step:approve) before executing any proposed skill creation or modification.
[SAFE]: File access is limited to reading the skill-index and workspace conversation history to identify patterns, which aligns with its stated purpose of gap analysis.
[SAFE]: No external network operations, remote script downloads, or unauthorized command execution patterns were detected.
[SAFE]: The self-improvement and next-skill protocols are configured to be non-autonomous, requiring the user to pick or approve recommendations.

insight-skill-gap