self-improving-agent

Log command failures and corrections for Minis by triggering the scripts/minis_auto_log.sh script, which records entries as organized learnings under /var/minis/skills/self-improving-agent/data. This centralizes and persists learning signals from errors and user corrections.

Persisting knowledge gaps and error corrections across projects is handled by centralizing learning signals into organized data structures. This approach captures command failures and applies them across Minis projects to enable a traceable improvement loop.

Yes, you can share logged learnings by promoting reusable entries from project-scoped environments to a public area. This supports clear separation of contexts and enables cross-project sharing via the Minis auto-log tooling.

Yes, the Minis auto-log script supports project-scoped logging and optional public logging, ensuring clear separation of contexts. Entries are stored under /var/minis/skills/self-improving-agent/data and /var/minis/skills/self-improving-agent/data/public respectively.

Centralizing learning signals helps with command failures and feature requests by capturing and organizing them as learnings, which creates a traceable improvement loop. This prevents recurring errors and addresses knowledge gaps within Minis.