ROCm/rocm-systems

amdsmi-changelog-automation

Check and generate changelog entries for amd-smi. Use when: reviewing PRs for changelog updates, generating release notes, checking CHANGELOG.md compliance.

amdsmi-commit-and-pr-conventions

Use when writing or restructuring git commits or opening/updating a pull request for amd-smi — composing commit titles, commit message bodies, PR titles, or PR descriptions. Defines the Conventional Commits `type(amdsmi):` house title convention, the…

thread-trace

Diagnoses why a compute kernel on an AMD GPU is slow, instruction by instruction, with Advanced Thread Trace (rocprofv3 --att) and the rocprof-trace-decoder Python API. Use when a kernel runs below its expected throughput and you need the instructions and…

decode-and-triage

Replays an HRR capture archive (a `.hrr` directory from HIP Record and Replay, or one of its `pid-*` process directories) on an AMD GPU and reports what failed: outcome, fault class, faulting kernel, fault address and failing event, without the original…

enable-recording

Records a failing HIP or ROCm GPU workload into an HRR archive that can be replayed and debugged later, and checks that archive is readable before anyone is asked to send it. Use when an AMD GPU workload crashes, hangs, aborts with a memory access fault, or…

rocprof-compute-kernel-bottleneck

Profiles an AMD GPU application with rocprof-compute and finds which kernel is slow and why, using performance metrics and PC sampling. Collects counters and selects kernels and dispatches, then analyzes GPU memory throughput, compute efficiency, wavefront…

rocprof-compute-memory

Analyzes the AMD GPU memory hierarchy with rocprof-compute, covering the Memory Chart, vL1D and L2 caches, LDS bank conflicts, Infinity Fabric traffic, and the experimental gfx950 memory bandwidth analysis. Use when the user asks about memory bandwidth, cache…

rocprof-compute-pc-sampling

Collects and analyzes rocprof-compute PC sampling to show which individual instructions in an AMD GPU kernel are hot and why wavefronts stall there, with ISA and source line attribution. Use when the user asks which instruction or source line is slow, asks…

rocprof-compute-roofline

Collects and reads rocprof-compute roofline data to place an AMD GPU kernel against the hardware's peak compute and peak bandwidth ceilings. Use when the user asks whether a kernel is compute-bound or memory-bound, asks about arithmetic intensity or FLOPs per…

rocprof-compute-speed-of-light

Reads the rocprof-compute System Speed-of-Light panel to show how close an AMD GPU kernel runs to hardware peak for compute throughput, memory bandwidth, and utilization. Use when the user asks how close to peak a kernel is, what percent of peak it reaches,…

rocprof-compute-torch-trace

Attributes AMD GPU kernel performance counters back to the PyTorch or Triton operators that launched them, using rocprof-compute torch and triton tracing. Use when the user asks which PyTorch layer or operator is slow, wants to map GPU kernels to model code,…

pc-sampling

Profiles a target application using rocprofv3 with program counter sampling enabled and then analyzes the results. Use when the user has a program that runs on an AMD GPU and asks to perform PC sampling, to determine the runtime performance characteristics of…

validate

Run full rocDecode build and test pipeline after code changes

rocprofv3-doctor

Diagnoses why rocprofv3 or ROCm GPU profiling does not work on a machine, by running `rocprofv3 --doctor` and turning its report into ordered, root-cause fixes. Use when rocprofv3 fails, crashes, hangs, finds no GPU, or will not start; when an error such as…

choosing-a-profiling-preset

Recommends which rocprofiler-systems runtime profiling preset (balanced, trace-gpu, trace-hpc, sys-trace, workload-trace, etc.) to pass to `rocprof-sys-run --preset=<name>` or `rocprof-sys-sample --preset=<name>` based on the user's workload and profiling…

discovering-profiling-options

Answers questions about what rocprofiler-systems configuration options exist and how to use them — e.g. "what hardware counters are available on my system", "what ROCm APIs can I trace", "what settings control sampling/output", "how do I persist my…

instrumenting-binaries

Guides instrumentation with `rocprof-sys-instrument` - choosing runtime instrumentation vs binary rewrite (`-o`), including or excluding functions and modules with regex options, previewing what will be instrumented with a dry run (`--simulate`,…

rocprofv3

Profiles applications on AMD GPUs with rocprofv3 and explains the results: application tracing (hot kernels, memory copies, idle GPU, Perfetto timelines), hardware counters (occupancy, cache hit rate, HBM bandwidth, memory- or compute-bound), Advanced Thread…

feature-unit-testing

Use when writing, planning, or improving unit tests for low-level transport or systems code — especially when reasoning about branch coverage, test gaps, identifying which uncovered paths are worth pursuing, or deciding when a feature's test suite is ready to…

amdsmi-ci-logs

Use when a GitHub Actions run or job for an amd-smi / rocm-systems PR failed and the web UI shows truncated output, when you need the full failing log, or when you need CI artifacts (test reports, packages). Triggers: an actions/runs/<id>/job/<id> URL, 'where…

amdsmi-test-runner

Run C++ and Python tests for amd-smi. Use when: running tests, verifying test results, checking test coverage, pre-review test validation.

amdsmi-using-git-worktrees

Use when starting amd-smi feature work, executing an implementation plan, or reviewing a PR that needs an isolated workspace away from the main checkout. Sets up a worktree following the rocm-systems-pr<PR#> convention.

amdsmi-improvement-evaluation

A structured approach to evaluating AMD SMI's current state and identifying improvement opportunities.

pressure-testing-skills

Use when validating or hardening a SKILL.md, prompt, rule, or any agent-followed document — when you need proof it actually works, not just that it reads well. Runs a fresh subagent that follows the doc literally on a known-answer fixture, captures friction,…

writing-skills

Use when creating or editing a SKILL.md file under .claude/skills/. Defines the TDD-for-documentation discipline, structure, and CSO rules that every amd-smi skill must follow.

perf-check

Run the rocDecode performance test and compare to baseline to catch regressions

amdsmi-restructure-commits

Use when finishing an amd-smi development branch — consolidating commits into logical groups with clean messages AND deciding how to integrate the work (merge to develop, push and open PR, keep as-is, or discard). Covers commit restructuring plus the…

writing-plans

Use when an approved spec exists and you need a bite-sized, file-level implementation plan before any code is written. Produces a plan ready for executing-plans or subagent dispatch.

rdc-build-install

Build and install RDC from source. Use when: building locally, installing before tests, pre-review build verification, build + install + verify. Requires GRPC_ROOT to be set.

amdsmi-agent-handoff

Use when handing work from one amd-smi agent to another (planning→development, planning→review, development→review) or compacting a long session into a fresh one. Produces a compact handoff doc referencing artifacts by path instead of duplicating them.

amdsmi-interrogate

Use whenever amd-smi work begins from a design — a Confluence Feature Design, a Jira/SWDEV ticket, a driver hand-off, or a user description. Reconciles the design against what the code actually does, questions it adversarially, and produces an approved spec.…

executing-plans

Use when a written implementation plan exists at ${TMPDIR:-/tmp}/amdsmi-agent-plans/ and you need to execute it task-by-task in the current session with verification at each step.

test-driven-development

Use when implementing any amd-smi feature, bug fix, or behavior change — before writing implementation code. Enforces strict RED-GREEN-REFACTOR: failing test first, watch it fail, minimal code to pass, refactor.

dispatching-parallel-agents

Use when facing 2+ independent problems with no shared state — e.g., unrelated test failures in different subsystems, multiple independent bug investigations, parallel research tasks. Dispatch one focused subagent per domain instead of investigating…

systematic-debugging

Use when encountering any bug, test failure, build failure, or unexpected behavior in amd-smi — before proposing any fix. Enforces root-cause investigation before symptom patching.

amdsmi-build-install

Build and install amd-smi from source. Use when: building locally, installing before tests, pre-review build verification, build + install + verify.

amdsmi-packaging-test

Integration testing for amd-smi Python packaging. Use when: testing package installs, multi-OS testing, manylinux wheel builds, offline install verification, glibc/Python version compatibility, post-install smoke tests.

amdsmi-output-verification

A checklist for verifying whether AMD SMI is reporting GPU information correctly.

amdsmi-python-style-guide

ROCm Python style guide based on TheRock. Use when: writing Python code, reviewing Python PRs, checking Python style, creating Python scripts, type hints, error handling patterns, pathlib usage, argparse CLI design.

verification-before-completion

Verification checklist before marking work complete. Use when: finishing a fix, completing a review finding, validating a build, confirming a refactor.