Browse reusable Agent Skills, each with a clear purpose and practical guidance.
Verify that a pull request into microsoft/vscode-cmake-tools meets contribution requirements. Use when preparing, reviewing, or finalizing a PR to check for a descriptive title, a meaningful description, a properly formatted CHANGELOG entry, code correctness, regression risks, adherence to existing patterns, and whether documentation updates are needed.
Shake down a flaky test (or subset of flaky tests) in dask/distributed. Reproduce locally with pytest-repeat, then either fix-and-verify locally or drive the `torture-test` GitHub Action via `gh` until CI is green. Use when the user says things like "debug flaky test <test name>", "fix intermittent failures in test <test name>", "torture test <test name>", "track down a flaky test", "reproduce a flaky failure", "fix intermittent test failures", or otherwise asks to shake out / harden a test that fails only sometimes.
Review UCX pull requests with the project review style, GitHub workflow, and reviewer checklist. Use when asked to review a UCX PR, inspect a GitHub pull request, draft review comments, triage PR scope, or apply REVIEW.md guidance.
Use when work in openclaw-channel-dingtalk involves DingTalk real-device validation, PR-scoped testing checklists, 验证 TODO drafting, or contributor guidance for this workflow.
Systematic methodology and concrete tool recipes for debugging Unix, Python, and PyTorch programs - crashes, hangs, segfaults, wrong output, CUDA OOM, NaN/Inf, slowness, and multi-node/multi-GPU issues. Use when a program crashes, hangs, deadlocks, segfaults, runs out of memory (OOM), produces NaN/Inf or wrong numbers, runs too slowly, or when the user mentions gdb, strace, py-spy, core files, CUDA_LAUNCH_BLOCKING, ldd/nm/LD_PRELOAD, cProfile, or distributed training hangs. Distilled from "The Art of Debugging", the latest version of which can be found at https://github.com/stas00/the-art-of-debugging The latest SKILL.md version can be found at https://github.com/stas00/the-art-of-debugging/blob/master/SKILL.md
Validate and analyze TensorRT performance data from paired layer-info JSON and profile/latency JSON files. Use when asked to inspect TensorRT, TRT, torch-tensorrt, or ONNX-TensorRT perf reports, verify that layer/profile JSON files are valid and from the same model, infer basic model information, find likely fusion or latency optimization opportunities, and produce a concise Markdown performance report or structured JSON data.
Run the ISR rolling-restart chaos drill for RobustMQ — 10 rounds of kill-one-node + write + restart + verify across a 3-replica shard. Use when the user wants to verify replica replication / ISR health, failover, or behaviour under churn — e.g. "演练 isr 副本同步", "验证三副本同步", "kill leader 演练", "验证 leader 切换/故障转移", "滚动重启/频繁切换 isr 是否正常", "test ISR replication", "副本同步是否正常".
Run a 3-node meta-service Raft cluster failover drill for RobustMQ. Use when the user wants to verify cluster startup, node removal/recovery, leader election, snapshot replication, or graceful shutdown — e.g. "演练 raft 集群", "test cluster failover", "验证节点删除/恢复", "run the raft drill".
Deep analysis and iterative fixing of a Rust source file. Finds logic errors, lock/concurrency issues, and simplification opportunities, then fixes them one by one until the file is clean.