skill-forge

Automate skill generation by detecting recurring sequences of tool usage and user interactions from session history, then codifying those patterns into new `SKILL.md` files. This transforms implicit workflow knowledge into explicit, reusable automation for future tasks.

Pattern detection for workflow codification identifies repeated sequences of actions across past sessions. It validates these extracted sequences against historical examples to ensure quality before automatically generating self-improving AI skills.

Turn session history into reusable AI skills by searching past memory for recurring user interactions and tool usage sequences. The extracted patterns are validated against previous examples and automatically compiled into structured `SKILL.md` files.

Yes, you can auto-generate a skill for a specific deployment and notification sequence. If you consistently use a specific sequence of commands for tasks like releasing updates, the pattern detection mechanism codifies it into a single automated skill.

No external dependencies are required to codify recurring tool usage into skills. The skill generation process relies entirely on internal memory search, pattern extraction, and validation against past examples to build self-improving automation.

Self-improving AI skill generation from patterns requires sufficient session history to accurately detect and validate recurring sequences. If past interactions lack consistent tool usage patterns, the automated codification into `SKILL.md` files may not produce relevant or high-quality skills.