Agent Skills by Junhyek Han
Showing 24 vetted skills indexed across 1 GitHub repositories.
junhyekhrlxp-reward-engineering
Design minimal reward logic patches for reinforcement-learning tasks.
junhyekhrl-task-audit
Audit reinforcement-learning repositories for task, commands, rewards, logging, and metrics.
junhyekhrlxp-monitor-training
Monitor active RL training runs for failures and metric regressions.
junhyekhrlxp-audit-task
Audit RL repositories to identify implemented tasks and generate task cards.
junhyekhrlxp-queue-batch
Summarize ready, running, completed, and blocked jobs from RLXP batch manifests.
junhyekhrlxp-analyze-run
Compare reinforcement-learning run metrics against an incumbent and classify the decision.
junhyekhrlxp-validate-candidate
Validates RLXP candidates against schema, contract scope, and launch safety constraints.
junhyekhrlxp-curriculum-design
Design curriculum adjustments for reinforcement learning tasks using per-bin evidence.
junhyekhrlxp-propose-candidates
Propose bounded RL experiment candidates from run evidence and contract constraints.
junhyekhrlxp-plan-parallel-batch
Plan parallel execution batches for validated RLXP candidates with seed and GPU assignment.
junhyekhrl-experiment-loop
Coordinate reinforcement-learning experiment loops with validation, monitoring, and launch gates.
junhyekhrlxp-make-contract
Create or update study-scoped RL experiment contracts from approved task definitions.
junhyekhrlxp-update-report
Update RLXP study reports from structured experiment evidence.
junhyekhrlxp-dr-design
Design domain randomization ranges and schedules for reinforcement learning experiments.
junhyekhrlxp-autoloop
Resume and govern an autonomous RLXP experiment loop from existing study state.
junhyekhrlxp-dr-probe
Plan one-parameter-at-a-time domain-randomization probes for incumbent reinforcement-learning policies.
junhyekhrlxp-plan-baseline
Create a study-scoped reinforcement-learning baseline plan with training commands and metric paths.
junhyekhrl-experiment-plan
Convert RL experiment ideas into scoped contracts with metrics and approval gates.
junhyekhrlxp-reward-review
Reviews reinforcement-learning reward changes for task alignment and approval readiness.
junhyekhrl-reward-curriculum-design
Route legacy RL reward and curriculum requests to RLXP intervention skills.
junhyekhrlxp-reward-reflection
Generate a reward reflection JSON with component diagnosis and next action.
junhyekhrlxp-define-metrics
Define task-scoped reinforcement-learning metrics, guardrails, and scorecards for a study.
junhyekhrlxp-audit-results
Audit reinforcement-learning improvement claims for validity across RLXP experiments.
junhyekhrlxp-reward-tuning
Adjust existing RL reward weights, thresholds, scales, or schedules within contract scope.