Skip to content

fix(evolution): keep skill contracts growth-safe #9

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
jhinpan merged 1 commit into from
Jul 23, 2026
Merged

fix(evolution): keep skill contracts growth-safe #9

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion SKILL.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
---
name: rocm-kernel-wiki
description: Search and apply ROCmKernelWiki when optimizing AMD Instinct kernels for MI300/gfx942 or MI350/MI355X/gfx950. Use for MFMA, LDS, direct-to-LDS, s_waitcnt, FP8/FP6/FP4/MXFP, wave reductions, GEMM/attention/MoE, CUDA-to-HIP migration, and CK/CK-Tile/AITER/hipBLASLt/FlyDSL/Triton/HIP/GCN implementations. Also use for merged-PR evidence from composable_kernel, AITER, Tensile, rocBLAS, flash-attention, FlyDSL, Triton, vLLM, or SGLang. Do not use for other GPU architectures, NVIDIA-only kernels, host framework behavior, or ROCm installation and driver troubleshooting.
description: Search and apply ROCmKernelWiki when optimizing AMD Instinct kernels for MI300/gfx942 or MI350/MI355X/gfx950. Use for MFMA, LDS, direct-to-LDS, s_waitcnt, FP8/FP6/FP4/MXFP, wave reductions, GEMM/attention/MoE, CUDA-to-HIP migration, and CK/CK-Tile/AITER/ATOM/hipBLASLt/FlyDSL/Triton/HIP/GCN implementations. Also use for merged-PR evidence from ROCm/rocm-libraries, AITER, ATOM, flash-attention, FlyDSL, Triton, vLLM, or SGLang. Do not use for other GPU architectures, NVIDIA-only kernels, host framework behavior, or ROCm installation and driver troubleshooting.
---

# ROCmKernelWiki — AMD CDNA Kernel Optimization Wiki
Expand Down
8 changes: 4 additions & 4 deletions tests/test_validate.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -119,9 +119,6 @@ def test_codex_skill_contract():
assert re.fullmatch(r"[a-z0-9]+(?:-[a-z0-9]+)*", skill["name"])
assert len(skill["description"]) <= 1024
assert "merged-PR" in skill["description"]
for repo in (ROOT / "sources/prs").iterdir():
if repo.is_dir():
assert repo.name.lower() in skill["description"].lower(), repo.name
assert len(skill_text.splitlines()) < 500
assert "~/.claude/skills" not in skill_text
assert "$HOME/.agents/skills" in readme
Expand Down Expand Up @@ -234,7 +231,10 @@ def test_get_page_scope_and_index():
try:
from _index import id_index

assert len(id_index()) == 7542
manifest = yaml.safe_load(
(ROOT / "data/corpus-manifest.yaml").read_text(encoding="utf-8")
)
assert len(id_index()) == manifest["counts"]["unique_page_ids"]
finally:
sys.path.remove(scripts_dir)

Expand Down
Loading