// Create a feature branch, commit changes, push, and open a PR against iopsystems/llm-perf
| name | pr |
| description | Create a feature branch, commit changes, push, and open a PR against iopsystems/llm-perf |
Create a PR from the current uncommitted changes. Handles branching, committing, pushing, and opening the PR against the upstream repo.
The skill accepts an optional branch name argument:
/pr fix-socket-perms)Verify prerequisites:
git status
git diff
git diff --staged
If there are no changes, stop and tell the user.
Analyze the changes:
git log --oneline -10
Create feature branch:
main, create a new branch:
git checkout -b <branch-name>
Stage and commit:
git add -A)type(scope): description)feat, fix, refactor, chore, docs, testCo-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> trailerPush to origin:
git push -u origin <branch-name>
Open PR against upstream:
iopsystems/llm-perforigin (determine the owner from git remote -v)gh pr create with --repo iopsystems/llm-perf and --head <fork-owner>:<branch-name>gh pr create \
--repo iopsystems/llm-perf \
--head <fork-owner>:<branch-name> \
--title "<conventional commit style title>" \
--body "$(cat <<'EOF'
## Summary
<1-3 bullet points describing the changes>
## Test plan
<bulleted checklist of testing done or needed>
Generated with [Claude Code](https://claude.com/claude-code)
EOF
)"
Report the PR URL to the user.
cargo clippy, cargo test, or cargo fmt --check haven't been run yet during this session, run them before committinggit@github.com:brayniac/llm-perf -> brayniac)Use when adding a new model (or model family) to the llm-calc database (calc/src/data/models.ts). Walks through sourcing fields from HuggingFace config.json, reading papers for novel attention variants, writing tests first, and registering the entry. Invoke this whenever a model is being added or updated — even for a single SKU, since the failure modes (wrong KV bytes, mis-sized attention, silently broken roofline) all look fine until someone runs the calc against a real workload.
Use when adding a new accelerator (GPU/TPU/Trainium/Gaudi), interconnect fabric, or multi-accelerator system to the llm-calc database (calc/src/data/{accelerators,interconnects,systems}.ts). Routes to the right sub-procedure based on whether the SKU is a new chip, a new fabric, or a new product composition. Invoke whenever hardware is being added — the failure modes (sparsity-inflated TFLOPS, confused per-direction vs aggregate BW, wrong variant in a system) all produce plausible-looking perf numbers that are silently wrong.
Create a release PR with version bump and changelog update