What problem does it solve?
This Skill systematizes agent improvement by turning vague performance issues into a repeatable adversarial training loop that diagnoses weaknesses, tests them with targeted challenges, and records confirmed learnings.
Core Features & Use Cases
- Weakness Diagnosis: Profiles an agent's recurring failures, knowledge gaps, and durable strengths from learnings history and project signals.
- Targeted Challenge Generation: Creates focused, grounded challenges that probe specific weaknesses with clear acceptance criteria and hidden traps.
- Serial Evaluation and Learning: Runs one challenge at a time, evaluates results and calibration, then updates learnings and capability tracking for future improvement.
- Use Case: Use this Skill when an agent keeps making the same kind of mistake and you want structured evidence-driven training instead of ad hoc prompting.
Quick Start
Run active-learn on the target agent or topic with a round count and let it generate, execute, evaluate, and persist the training cycle automatically.
