hf-release-notes
Generate Hugging Face Hub (huggingface_hub) release notes from cached PR JSON files. Use when asked to draft release notes from PR files.
Works with
---
name: hf-release-notes
description: Generate Hugging Face Hub (huggingface_hub) release notes from cached PR JSON files. Use when asked to draft release notes from PR files.
license: Apache-2.0
---
# HF Release Notes
## Overview
Generate release notes for huggingface_hub from cached PR JSON files. This skill reads PR metadata, categorizes entries, and produces a formatted markdown release notes document.
**Output directory:** The prompt will specify the output directory as `<output_dir>`. All paths below use this placeholder.
## Workflow
### 1. Read PR data
Read all PR JSON files from `<output_dir>/tmp/pr_*.json`. Each file contains:
```json
{
"number": 1234,
"title": "...",
"author": "username",
"merged_at": "2026-01-15T10:30:00Z",
"body": "...",
"labels": ["highlight", "cli"],
"url": "https://github.com/huggingface/huggingface_hub/pull/1234",
"doc_diffs": [
{
"filename": "docs/source/en/guides/cli.md",
"status": "modified",
"patch": "@@ -10,6 +10,10 @@ ..."
}
]
}
```
The `doc_diffs` field contains unified diffs for any `.md` files under `docs/` that were
changed in the PR. This is empty (`[]`) for PRs with no documentation changes.
### 2. Identify highlights
A PR should be highlighted if:
- It has the `"highlight"` label on GitHub, **or**
- You judge it significant enough to deserve a detailed section (e.g., a major new
feature, a meaningful UX improvement, or a notable breaking change) even without the label.
Use your judgment — not every feature PR needs a highlight, but don't limit highlights
to only labeled PRs. Each highlight section follows this structure:
1. **Emoji header** — e.g., `## 🖥️ New CLI commands`
2. **Prose summary** — 2-5 sentences describing the user-visible change in flowing text.
Avoid bullet points here; write natural sentences. Only use bullets if the content
truly calls for a list (e.g., enumerating 4+ distinct sub-features).
3. **Code examples** — if the PR introduces new commands or APIs, include a fenced code
block with a concrete usage example (cherry-pick from doc diffs when available).
4. **PR attribution lines** — one bullet per PR that contributed to this highlight:
`- PR title by @author in #1234`
### 3. Classify standard items
For non-highlight PRs, classify into sections using `references/sections.md` heuristics based on:
- PR labels
- PR title keywords
- PR body content
### 4. Use doc diffs and fetch relevant documentation
For highlighted PRs and other PRs that introduce new features, commands, or APIs:
**Use doc diffs first:**
- Check the `doc_diffs` field in the PR JSON. If present, these contain the actual
documentation changes made in the PR (unified diff format).
- Use these diffs to understand what was documented, extract code examples, and write
more accurate summaries. The diffs show exactly what the PR author wrote in the docs.
- The `filename` field maps to a docs page URL. For example,
`docs/source/en/guides/cli.md` → `https://huggingface.co/docs/huggingface_hub/main/en/guides/cli`
**Fetch full doc pages when needed:**
1. Start by fetching the docs index page to discover the site structure:
`https://huggingface.co/docs/huggingface_hub/main/en/index`
2. Based on the PR content (title, body, labels), identify which doc pages might be
relevant. For example:
- A CLI PR → check `https://huggingface.co/docs/huggingface_hub/main/en/guides/cli`
- An Inference PR → check `https://huggingface.co/docs/huggingface_hub/main/en/guides/inference`
- A new API class or method → check the corresponding reference page
3. Fetch the candidate doc page to confirm it covers the feature from the PR.
4. Only include a doc link if the page genuinely documents the feature.
**How to reference:**
- In highlight sections, add a line like:
`📚 **Documentation:** [Guide name](https://huggingface.co/docs/huggingface_hub/main/en/guides/...)`
- In standard sections, append the doc link inline after the attribution when relevant:
`- PR title by @author in #1234 — [docs](https://huggingface.co/docs/huggingface_hub/main/en/...)`
- Do NOT add doc links for internal/CI/test PRs or PRs that don't have user-facing docs.
- Do NOT fabricate doc URLs — only link to pages you have actually fetched and verified.
### 5. Generate release notes
Output to `<output_dir>/RELEASE_NOTES_<version>.md` using the structure from `references/release-notes-template.md`:
- Title: `# [vX.Y.Z] <tagline>` (derive tagline from main highlights; always use the base version without prerelease suffix, e.g. `[v1.8.0]` even when generating for `v1.8.0.rc0`)
- One section per highlight with emoji header and narrative
- Standard sections for remaining items (only include sections with items)
### 6. Quality checks
Before finishing:
- Verify every PR from `<output_dir>/tmp/` appears exactly once
- **Do NOT include any PR that does not have a corresponding file in `<output_dir>/tmp/`** — those PRs belong to a different release and must not appear in these notes
- No empty sections
- Consistent emoji headings
- Every item ends with attribution: `by @author in #1234`
- Doc links point to real, verified pages
## Input
- Version string (e.g., "v1.3.8")
- PR JSON files in `<output_dir>/tmp/`
## Output
- Release notes markdown at `<output_dir>/RELEASE_NOTES_<version>.md`
## Resources
- `references/release-notes-template.md`: Skeleton structure for release notes
- `references/sections.md`: Keyword-based section mapping and guidance
- Documentation site: `https://huggingface.co/docs/huggingface_hub/main/en/index`More Writing & Documentation skills
paper-context-resolver
lllllllama/rigorpilot-skills
Rigor Paper Context helper for README-first deep learning repo reproduction. Use only when the README and repository files leave a narrow reproduction-critical gap and the task is to resolve a specific paper detail such as dataset split, preprocessing, evaluation protocol, checkpoint mapping, or runtime assumption from primary paper sources while recording conflicts. Do not use for general paper summary, repo scanning, environment setup, command execution, title-only paper lookup, or replacing README guidance by default.
repo-intake-and-plan
lllllllama/rigorpilot-skills
Rigor Intake helper for README-first deep learning repo reproduction. Use when the task is specifically to scan a repository, read the README and common project files, extract documented commands, classify inference, evaluation, and training candidates, and return the smallest trustworthy reproduction plan to the main orchestrator. Do not use for environment setup, asset download, command execution, final reporting, paper lookup, or end-to-end orchestration.
minimal-run-and-audit
lllllllama/rigorpilot-skills
Rigor Run skill for README-first deep learning repo reproduction. Use when the task is specifically to capture or normalize evidence from the selected smoke test or documented inference or evaluation command and write standardized `repro_outputs/` files, including patch notes when repository files changed. Do not use for training execution, initial repo intake, generic environment setup, paper lookup, target selection, hidden scientific-meaning changes, or end-to-end orchestration by itself.

