Menu
This skill should be used when the user wants to set up checkpoints, record approval decisions, gate execution pending human confirmation, or manage human-in-the-loop review points in a computational math workflow.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Route AI4Math computational mathematics tasks to the appropriate bundled skill package.
Use when reproducing, deploying, diagnosing, tuning, visualizing, or reporting computational math research code with a coding agent
This skill should be used when the user asks to search for external algorithms, find related implementations, discover alternative approaches, or look up papers and code for computational math optimization methods.
This skill should be used when the user asks to tune parameters, run hyperparameter search, optimize algorithm settings, or find the best configuration for a computational math method.
Use when classifying a computational math repository or task into a broad domain before choosing specialist Skills, references, runtimes, and validation evidence.
Use when a user starts an end-to-end computational math research-code reproduction, deployment, tuning, visualization, or reporting workflow with a coding agent.
| name | human-review-skill |
| description | This skill should be used when the user wants to set up checkpoints, record approval decisions, gate execution pending human confirmation, or manage human-in-the-loop review points in a computational math workflow. |
Manages lightweight human approval gates for the compact workflow.
For end-to-end computational math research-code reproduction workflows, this Skill should be selected by computational_math_reproduction_workflow_skill only when durable approval logs are needed.
Four approval types, all conversation-led:
plan.md)repair_plan.md)tuning/tuning_plan.md)User replies approve, ok, yes, 可以, 好的, 继续, 同意, 批准, 就这个吧, 选这个, 没问题 are all treated as approval.
Revise: revise, 修改, 改一下
Reject: reject, 拒绝, 不行
Skip: skip, 跳过, 跳过这个
plan.md, repair_plan.md, tuning/tuning_plan.md) to the user.approve, revise, reject, or skip.approvals/approval_log.jsonl when the operation is high-risk (source changes, dependency installs, budget increases).01_task_understanding.md, etc.) by default.Input: --run <path>, --checkpoint <name>, --risk-level <low|high>
Exits 0 if allowed, exits 2 if blocked.
Records an approval decision to approvals/approval_log.jsonl. Only call when explicitly logging a high-risk approval.
python -m skills.human_review_skill.scripts.approval_logger --run /path/to/output --checkpoint plan --decision approve --reason "low-risk demo"
approve, revise, reject, or skip.Do not write numbered checkpoint files by default. Process reasoning stays in conversation.