dev-learner

To identify repeated conversation workflows, you need pattern detection to analyze your session and flag sequences occurring 3 or more times. The dev-learner meta-skill automates this by extracting triggers, steps, and involved agents to propose reusable skills.

Yes, you can extract reusable skills from engineering retrospectives automatically. The dev-learner skill analyzes these sessions to detect recurring multi-agent handoffs and outputs, then generates a structured learner report proposing formalized skills.

The best way to reduce repeated manual prompting in multi-agent workflows is detecting recurring sequences and converting them into structured proposals. This approach extracts defined inputs, outputs, and steps, creating a documented handoff plan to formalize the pattern.

The dev-learner skill detects patterns and proposes new skills, but does not formalize them directly. It produces a structured learner report and an optional handoff plan to a skill-creator, which handles the actual creation of the reusable skill.

To document repeating multi-agent handoffs, you apply pattern detection to extract the triggers, involved agents, and defined inputs/outputs. This generates a structured learner report detailing the recurring sequence for future skill creation.

The skill extraction process is triggered when a specific workflow, prompt sequence, or multi-agent handoff repeats 3 or more times during a development session. Pattern detection then flags this recurrence for structured extraction and proposal generation.