---
name: rocm-hipdnn-deep-review
description: "Multi-agent deep review of a hipDNN or dnn-providers PR, branch or commit range. Scouts map each affected area, focused reviewer lanes run in parallel, every finding is verified by source proof, reproduction or mutation, and a ledger carries findings across rounds until no blocking or important one stays open. Its IMPACT lane reports kernel ingestor changes that break at run time while the build stays green (kernel ABI, dispatch, descriptors) with the scripts of hipdnn-ingestor-impact, which ans"
metadata:
  decimalai-skill-id: "ff641e4a-b7d7-4313-a6ff-d87f8a2b8baa"
  decimalai-version-id: "9b3b927f-1029-48e1-850a-7faa561c9b6a"
  decimalai-version: "2"
  decimalai-body-sha256: "f24a66bc8b66ca4240a208a057e7e5b9573289d69ea46ecdc8773acf18c04c59"
  decimalai-name: "rocm/hipdnn-deep-review"
  decimalai-format: "1"
  decimalai-invocation: "model"
  decimalai-category: "code-review"
  decimalai-source: "https://app.decimal.ai/s/rocm-hipdnn-deep-review@2/SKILL.md"
---
# hipDNN Deep Review

`$SKILL` is the directory this skill was loaded from (the directory containing this
`SKILL.md`). You are the orchestrator: execute [RUNBOOK.md](RUNBOOK.md). You pin the change, dispatch
scouts, lanes and verifiers, merge their output and write the report; you do not review
code. Your formats are in [report.md](report.md); [RUNBOOK.md](RUNBOOK.md) *Fan out* says
what each agent reads, and `lanes/` holds the briefs. `RUN` builds with
[hipdnn-superbuild](../hipdnn-superbuild/SKILL.md) and tests with
[hipdnn-superbuild-test](../hipdnn-superbuild-test/SKILL.md); `IMPACT` runs the scripts of
[hipdnn-ingestor-impact](../hipdnn-ingestor-impact/SKILL.md). Install those skills alongside this one.

Prerequisites: network access, an authenticated `gh` with read access, `git` 2.38 or later (`pin.py` uses `git merge-tree --write-tree`), and Python 3 (it runs the scripts under
`$SKILL/scripts/`); `pre-commit` and `pytest` for the lanes that run validators. The `IMPACT` lane needs
Python 3.10 or newer and PyYAML; its emitter also needs Jinja2 ([hipdnn-ingestor-impact](../hipdnn-ingestor-impact/SKILL.md) *Prerequisites*). `RUN` and verifiers also need
a ROCm install (Linux: `/opt/rocm` or `ROCM_PATH`; Windows: the wheel SDK that [hipdnn-superbuild](../hipdnn-superbuild/SKILL.md) provisions), CMake, Ninja and about 25 GB free in `$SCRATCH`; without them `defect` and
`test-gap` findings stay `UNVERIFIED` and the review closes only through human
`ACCEPTED-RISK` decisions, recorded with `ledger.py set` (`PIN.md` records `Build host`).

Platforms: Linux, macOS and Windows. On Windows run the skill from Git Bash (Git for Windows): the runbook's shell blocks need
`bash`, `awk`, `sed`, `grep` and `comm`. Blocks write the interpreter as `python3`; on Windows that name is the Microsoft Store
stub, so use `python` or `py -3` instead. Scripts read and write UTF-8 and LF whatever the host's locale.

For a quick answer without a review, use [hipdnn-ingestor-impact](../hipdnn-ingestor-impact/SKILL.md): it runs the
same scripts as the `IMPACT` lane ([lanes/runtime-impact.md](lanes/runtime-impact.md)) and ends in a `Runtime impact:` line.

## Entry contract

Ask for what is missing; record it in `PIN.md`:

| Input | Notes |
|---|---|
| Target | PR URL or number, `branch:<name>`, or `<base>..<head>`; `intent:<file>` (what it is for, how tested, `Known issues`; required without a PR) holds the intent document, whose `Known issues` section lists open items the author already knows, numbered `K1`, `K2` and so on, and never renumbered; without a PR, M4 and M5 are `N/A (no PR)` ([policy.md](policy.md) *M-rules*) |
| `evidence:<dir>`, `scratch:<dir>`, `clone:<dir>` | `$EVIDENCE` outside the product tree; disposable `$SCRATCH`; optional clone (default `$SCRATCH/clone`) |
| `terms:<file>`, `knowledge:<dir>` | Optional private terms, one fixed string per line, outside the repository; optional `$KNOWLEDGE` store |
| `round:<N>` | Default: one more than the highest `$EVIDENCE/r<k>/`; 1 when none |
| `prior:<dir>` | Optional published record of earlier rounds that `$EVIDENCE` lacks (`LEDGER.md` and `r<k>/`); [RUNBOOK.md](RUNBOOK.md) *Iterate* seeds `$EVIDENCE` from it |
| Agent limits | Ask the user or the host for `Agents:` (maximum concurrent agents), `Cores:` (CPU cores for builds and tests) and `Mutation:` (a host recipe for mutating a built tree, or none); the skill assumes no values. They go into `dispatch-context.md` ([RUNBOOK.md](RUNBOOK.md) *Fan out*) |
| Scope notes | Ask the user for scope notes the change's description does not state, such as areas being deprecated, or `none`. Free text; it goes into `dispatch-context.md` as `Scope:` and is informational unless it says "do not file findings in <area>" ([lanes/common.md](lanes/common.md) *Rules*) |

Local builds and host-only tests are allowed. GPU, remote or shared compute, and every write
to GitHub, Jira or a remote branch need the user's approval for that exact action. The
review is advisory: the user posts the draft comments. Write only `$EVIDENCE`, `$KNOWLEDGE`
and `$SCRATCH`, never the product tree.

## Completion

A round is complete when its gates hold and `REPORT.md` and `LEDGER.md` exist for the head.
The review is closed (`Closure: CLOSED`) when no `BLOCKING` or `IMPORTANT` finding is `OPEN`
or `UNVERIFIED` and no `STOP` or decision is unresolved. Otherwise `Closure: OPEN` lists what
blocks it and *Iterate* starts the next round; the user decides when to stop.