菜单
Multi-agent code review. Use when the user explicitly requests independent verification, cross-agent/council review, or a release go/no-go or requested ML pre-training gate needs several independent reviewers. High-risk changes alone do not require this skill.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Shared multi-agent harness state and coordination for Codex, Claude Code, and Antigravity. Use for repo resume/state, shared memory or active-task handoff, plan claims/recovery, fail-ledger checks, stale `.oms` cleanup, role/executor setup, local provider calls or write delegation, patch admission/landing, artifact provenance, and prior-session handoff.
Discovery-focused chem-bio ML guardrails for small molecules, molecular 3D and physics, proteins and variants, biomolecular interactions, peptides and antibodies, reactions and synthesis, generative design, cellular omics and phenotypic profiling, and biomedical networks. Use before writing, reviewing, evaluating, or training on RDKit/SMILES/SELFIES, QSAR/ADMET, quantum or force data, MMseqs2/MSA/protein language models, PDB/AlphaFold/DMS, docking/DTI/PPI, peptide-MHC or antibody data, reaction/retrosynthesis/yield data, molecular or protein generation, RNA/oligonucleotide/CRISPR, single-cell perturbation/Cell Painting, or biomedical knowledge graphs. Enforce domain splits, leakage checks, label semantics, baselines, metrics, calibration, applicability domain, and provenance.
PyTorch training correctness for optimizers and schedules, distributed/DDP, variable-size masked losses, checkpoint save/load, CUDA runtime settings, and equivariant model checks. Use when writing or reviewing training loops, optimizer/scheduler setup, multi-GPU code, loss reduction, or checkpoints.
Hypothesis-driven loop for ML and scientific experiments. Use to turn a question into a falsifiable hypothesis, design the smallest disconfirming run, pre-register prediction/baseline/metric, interpret results, coordinate experiment claims, reproduce or compare runs, or trace an output to its run.
Slurm/HPC workflow helper. Use when working on clusters, partitions, nodes, sbatch/srun jobs, GPU/CPU allocations, queues, logs, checkpoints, or any resource-heavy command — e.g. "submit the job", "check the queue", "잡 돌려줘", "큐 확인", "GPU 몇 장", "학습 돌려줘" on a cluster; also digesting a long job log (job-digest, "로그 요약") and reconciling finished Slurm jobs into shared state (run-reconcile, sacct). Reads local cluster reference when generated.
Specification-first workflow for new projects, existing-repo onboarding, unresolved draft specs, and broad or architecture-shaping features. Detect project state before asking questions and resolve only choices that affect implementation. Clear bounded changes do not require an interview.
| name | peer-review |
| description | Multi-agent code review. Use when the user explicitly requests independent verification, cross-agent/council review, or a release go/no-go or requested ML pre-training gate needs several independent reviewers. High-risk changes alone do not require this skill. |
Get independent findings without outsourcing judgment. The parent reviews the evidence and owns fixes, verification, landing, commit, and release.
--verify CMD as a mechanical backstop. Reviewer consensus cannot turn a
failing command green.oms agent-run --mode read when three independent
signals are unnecessary.oms peer-review --repo . --prompt "Review this diff for blocking findings."
oms peer-review --repo . --base origin/main --verify "bash scripts/check.sh"
oms peer-review --repo . --ml # only for an explicit ML gate
Never send secrets, credentials, private keys, env files, private paths, machine/cluster details, raw datasets, checkpoints, or generated scratch state. If sensitive-looking content occurs inside an otherwise relevant diff, keep that portion local instead of line-redacting it into external context.
Local Codex, Claude Code, and Antigravity review CLIs may receive sanitized task context when this skill is explicitly invoked. This does not authorize installation, authentication, write mode, permission bypass, destructive commands, connectors, commits, or pushes.
Summarize findings first by severity with file/line evidence. Separate reviewer claims from parent-verified facts. Report provider failures/skips and the local verification result. State the final parent decision explicitly.