Parse session logs to generate atomic patches for agent memory, settings, and rules.
npx skills add https://github.com/mishahanin/heading-os --skill calibrate-mishahanin
Please help me install this Agent Skill. Skill: calibrate Source: https://github.com/mishahanin/heading-os/tree/main/.claude/skills/calibrate Command: npx skills add https://github.com/mishahanin/heading-os --skill calibrate-mishahanin
This skill addresses the drift between your evolving preferences and the AI's current behavior by capturing session-end learnings and applying them as structural improvements to your workspace.
Run the calibrate skill at the end of your session to capture and apply improvements to your agent configuration.
High-intent search queries and answers about installing and using this skill.
To automate AI agent self-improvement, you run the calibrate skill at the end of your session. It reflects on session logs to identify behavioral corrections and applies those learnings as patches to your agent's memory and rules.
The best way to update agent memory and settings automatically is through atomic patching. The skill proposes numbered patches to your configurations and applies them in a single git commit to capture structural improvements.
Yes, you need git and python3 installed in your environment to apply behavioral patches. These dependencies are required to parse session logs, generate the configuration patches, and commit the updates to your project.
You can safely apply automated workflow optimization because the skill uses safety-first routing. It automatically routes corporate-sensitive changes to a manual review queue, ensuring only safe updates are auto-applied.
AI agent behavior drifts from your preferences because current configurations do not capture session-end learnings. The calibrate skill solves this by reflecting on your recent interactions and applying structural improvements to your workspace.
F0WCUSAggregate historical agent memory files into summaries of decisions and open tasks.
shekerkammaAutomate git-based code shipping with tests, versioning, and pull requests.
ethan-rioExplain unfamiliar code sections with higher-level project context and module maps.
labrinyangRecord session mistakes, lessons, and patterns to .claude.md.
