mindful-hermes-continuous-improvement

What problem does it solve?

This skill prevents Hermes from merely recording errors and instead drives verified improvements by capturing regressions, root causes, fixes, and confirmation that the fix holds.

Core Features & Use Cases

• Improvement loop that enforces capture → fix → verify: keeps the system honest so learning becomes behavior change, not journaling theater.
• Regression, capability gap, and decision-miner review routing: classifies recurring signals (regressions, missing capabilities, decision candidates, janitor findings, and cron health) and routes them into the right review artifacts.
• Skill patching and inspection-first governance: patches governing Hermes skills when reusable fixes are clear, and uses inspection rather than memory alone to validate outcomes.
• Weekly review card and improvement-loop ownership: produces compact weekly inspection outputs to show whether Hermes is truly improving.

Quick Start

Use mindful-hermes-continuous-improvement when a correction reveals a repeatable Hermes behavior gap, and ask it to capture the issue, propose and apply the governing fix, then verify the fix is still working.