| 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. |
Human Review Skill
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.
Lightweight Approval Types
Four approval types, all conversation-led:
- plan approval — before executing external code (
plan.md)
- repair approval — before modifying source, dependencies, entrypoint, or data (
repair_plan.md)
- tuning approval — before starting or expanding tuning budget (
tuning/tuning_plan.md)
- final acknowledgment — before accepting conclusions
Approval Aliases
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
Workflow
- Present the compact artifact (
plan.md, repair_plan.md, tuning/tuning_plan.md) to the user.
- Ask for exactly one decision:
approve, revise, reject, or skip.
- Record the decision in conversation.
- Only write
approvals/approval_log.jsonl when the operation is high-risk (source changes, dependency installs, budget increases).
- Do not write numbered checkpoint files (
01_task_understanding.md, etc.) by default.
Scripts
approval_gate
Input: --run <path>, --checkpoint <name>, --risk-level <low|high>
Exits 0 if allowed, exits 2 if blocked.
approval_logger
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"
Agent Conversation Semantics
- Ask for exactly one decision:
approve, revise, reject, or skip.
- If the user approves, continue to the next stage.
- If the user asks to revise, keep the same artifact pending and address the requested changes.
- If the user rejects, record the rejection and stop or reroute.
- If the user skips, continue only if the stage is optional.
Do not write numbered checkpoint files by default. Process reasoning stays in conversation.
