New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Publish Tracelab Huggingface

by uw-syfi in uw-syfi/TraceLab

Prepare, publish, refresh, or validate the TraceLab public dataset on Hugging Face under UW-SyFI/TraceLab.

Apache-2.0Auto-check passedAI & LLM Engineering

Install Publish Tracelab Huggingface

skills CLI
$ npx skills add uw-syfi/TraceLab --skill publish-tracelab-huggingface -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install uw-syfi/TraceLab publish-tracelab-huggingface --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/uw-syfi/TraceLab.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.codex/skills/publish-tracelab-huggingface .claude/skills/publish-tracelab-huggingface && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
publish-tracelab-huggingface
GitHub stars
142
Token cost
~1.4k tokens
SKILL.md length
474 words
Files
5 (incl. scripts)
Skills in repo
7
Repo updated
First seen
Licence
Apache-2.0

At a glance

Prepare, publish, refresh, or validate the TraceLab public dataset on Hugging Face under UW-SyFI/TraceLab.

  • Works in 5 steps: Audit → Build staging → Validate locally → …
  • Hugging Face releases
  • SKILL.md covers Invariants, 1. Audit, 2. Build staging and 3. Validate locally, plus 4 more sections
  • Runs Python scripts from its folder; calls uv, gh and git; needs HF_TOKEN

What it does

Publish Tracelab Huggingface is an agent skill from uw-syfi/TraceLab. Prepare, publish, refresh, or validate the TraceLab public dataset on Hugging Face under UW-SyFI/TraceLab. Use for Hugging Face releases, Dataset Card updates, Viewer/Data Studio failures, version tags, loaddataset compatibility, or mirroring a GitHub TraceLab release while preserving exact JSONL and DuckDB assets.

Its SKILL.md is about 1.4k tokens, which your agent loads only when the skill is triggered. The skill folder holds 6 other files, including scripts (for example `agents/openai.yaml`, `scripts/export_parquet.py` and `scripts/validate_hub_release.py`).

It sits in AI & LLM Engineering, covering Model hubs and datasets. It works with Hugging Face, DuckDB, GitHub and Python. The repository describes itself as: An open toolkit and public dataset hub for collecting, sanitizing, analyzing, and visualizing coding agent traces. The licence is Apache-2.0.

When your agent uses it

  • Hugging Face releases
  • Dataset Card updates
  • Viewer/Data Studio failures
  • Loaddataset compatibility

Example prompts

  • “/publish-tracelab-huggingface”

Requirements

  • Python 3

Workflow steps

5 steps, taken from the step headings in SKILL.md.

  1. Audit
  2. Build staging
  3. Validate locally
  4. Upload
  5. Validate Hub and tag

What it can do on your machine

Read from SKILL.md and the folder at commit 11b8b14. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 3 files in scripts/ (Python), which the agent can run.

    Shell commands in SKILL.md call:

    • uv
    • gh
    • git
    • hf

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use uv, gh and git, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names these keys or tokens, usually read from environment variables:

    • HF_TOKEN

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Publish Tracelab Huggingface loads about 1.4k tokens when it runs. Until then it costs about 87 tokens; SKILL.md has 474 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~87
When it runs · the whole SKILL.md, loaded when a task matches
~1.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from uw-syfi/TraceLab at commit 11b8b14, republished under its Apache-2.0 licence (© uw-syfi). 474 words, ~1,361 tokens.

Download SKILL.mdSave it as .claude/skills/publish-tracelab-huggingface/SKILL.md (or your agent's skills folder). This skill also uses 4 other files; get the full folder from GitHub.
name
publish-tracelab-huggingface
description
Prepare, publish, refresh, or validate the TraceLab public dataset on Hugging Face under UW-SyFI/TraceLab. Use for Hugging Face releases, Dataset Card updates, Viewer/Data Studio failures, version tags, load_dataset compatibility, or mirroring a GitHub TraceLab release while preserving exact JSONL and DuckDB assets.

Publish TraceLab to Hugging Face

Publish exact GitHub release assets plus Viewer-friendly relational Parquet. Treat the release JSONL and DuckDB as provenance authorities; never rewrite them to satisfy Hugging Face.

Invariants

  • Resolve the exact repo, release tag, JSONL, DuckDB, hashes, and row counts first.
  • Keep syfi_coding_trace.jsonl.gz and syfi_coding_trace.duckdb byte-identical to GitHub.
  • Derive rounds, tool_calls, and timing_events Parquet only from that DuckDB.
  • Make rounds the default config; join child configs through round_pk.
  • Stage under $TMPDIR; preserve unrelated dirty files.
  • Use uv run python; never system Python or pip.
  • Authenticate with explicit HF_TOKEN; never overwrite a shared HF_HOME login.
  • Create/move the version tag only after the final Viewer-ready commit passes.

1. Audit

bash
gh release view VERSION --repo uw-syfi/TraceLab --json tagName,assets,body,url
sha256sum trace/syfi_coding_trace.jsonl.gz trace/syfi_coding_trace.duckdb
gzip -t trace/syfi_coding_trace.jsonl.gz
git status --short --branch

Confirm HfApi().whoami() is the intended account with write/admin access to UW-SyFI. Set a task-local HF_HOME under $TMPDIR so Xet logs do not use the shared cache; HF_TOKEN supplies auth.

2. Build staging

bash
release_staging_directory="$(mktemp -d "$TMPDIR/tracelab-hf.XXXXXX")"
uv run python .codex/skills/publish-tracelab-huggingface/scripts/export_parquet.py \
  --database trace/syfi_coding_trace.duckdb \
  --output-directory "$release_staging_directory/data/VERSION"

The exporter converts rounds.turn_id UUID to string because datasets rejects Arrow UUID, and casts tool_calls.command_exit_code HUGEINT to int64.

Copy exact JSONL/DuckDB assets into versioned auxiliary paths. Copy LICENSE-DATASET.md and NOTICE; create SHA256SUMS for all data files.

The Dataset Card must configure only stable Parquet:

yaml
configs:
  - config_name: default
    default: true
    data_files: [{split: train, path: data/VERSION/rounds/train.parquet}]
  - config_name: tool_calls
    data_files: [{split: train, path: data/VERSION/tool_calls/train.parquet}]
  - config_name: timing_events
    data_files: [{split: train, path: data/VERSION/timing_events/train.parquet}]

Keep nested JSONL outside configs: optional provider fields such as turn_id make the HF JSON builder freeze an incomplete early schema and later fail with column names don't match.

Document counts, schemas, sanitization, responsible use, CC BY 4.0, hashes, GitHub release, project site, load_dataset() examples, arXiv URL, and issue #22's conservative replay guidance. Clearly label Parquet as derived and JSONL/DuckDB as byte-identical originals.

3. Validate locally

bash
uv run python .codex/skills/publish-tracelab-huggingface/scripts/validate_local_release.py \
  --database trace/syfi_coding_trace.duckdb \
  --staging-directory "$release_staging_directory" \
  --version VERSION

Load all three configs locally with uv run --with datasets python before uploading.

Show full SKILL.md (209 more words)Show less

4. Upload

Create UW-SyFI/TraceLab private, upload staging with hf upload --repo-type dataset, then read back repo_info(..., files_metadata=True) and compare every LFS SHA-256. If public publication was requested and checks pass, set private=False.

Private Viewer requires PRO/Enterprise; a free org returns 501. In that case validate locally first, publish, then immediately validate the public repo.

5. Validate Hub and tag

Run the Hub validator in tmux because first indexing/full loading can take minutes:

bash
TMUX_TMPDIR="$TMPDIR" tmux new-session -d -s tracelab-hf-validate -c "$(pwd)" \
  "uv run --with datasets python \
  .codex/skills/publish-tracelab-huggingface/scripts/validate_hub_release.py \
  --repo-id UW-SyFI/TraceLab --load-dataset \
  > '$TMPDIR/tracelab-hf-validate.log' 2>&1"

Require Viewer preview/search/filter, three Parquet configs with no pending/failed jobs, exact Hub row counts, and exact load_dataset() counts. statistics=false alone is non-blocking when Viewer works; HF can fail its histogram on constant-length pseudonymous session_id values.

Only then create the version tag at the exact final commit and verify the tag target.

Failure routing

  • column names don't match: JSONL is configured; switch to stable Parquet.
  • extension<arrow.uuid>: cast turn_id to string.
  • command_exit_code becomes float: cast HUGEINT to BIGINT.
  • private Viewer 501: locally validate, publish, then check public Viewer.
  • Xet shared-cache permission error: use task-local HF_HOME plus HF_TOKEN.
  • response is not ready yet: poll; pending is not failure.
  • failed config: inspect the Dataset page embedded traceback before changing data.

Scripts

  • scripts/export_parquet.py: export compatible relational Parquet.
  • scripts/validate_local_release.py: verify checksums, gzip, counts, and Parquet types.
  • scripts/validate_hub_release.py: poll Viewer/Parquet/size and optionally run load_dataset.

© uw-syfi, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 4 other files (scripts) in .codex/skills/publish-tracelab-huggingface of uw-syfi/TraceLab.

  • SKILL.md
  • agents/openai.yaml
  • scripts/export_parquet.py
  • scripts/validate_hub_release.py
  • scripts/validate_local_release.py

Open the folder on GitHubat commit 11b8b14

Compare with similar skills

Publish Tracelab Huggingface next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Publish Tracelab Huggingface compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Publish Tracelab Huggingface this skilluw-syfi/TraceLab142—~1.4kAutomated safety check: PassApache-2.0
DB AI GitHub Paper Weekly Newsdigoal/blog8.6k—~1.2kAutomated safety check: PassGPL-2.0
DB AI GitHub Paper Weekly Newsdigoal/blog8.6k—~2.4kAutomated safety check: PassGPL-2.0
Esmfold2JimLiu/science-skills2284 repos~2.5kAutomated safety check: PassApache-2.0
Qwen Mtp GgufR6410418/Jackrong-llm-finetuning-guide1.7k—~1.7kAutomated safety check: PassMIT
Hugging Face TokenizersOrchestra-Research/AI-Research-SKILLs13k6 repos~3.4kAutomated safety check: PassMIT

Similar skills

  • DB AI GitHub Paper Weekly News

    digoal/blog

    抓取并汇总"数据库、AI、GitHub、AI 论文"近 1 周新闻，输出图文并茂（含 mermaid 图）的 markdown 周报到当前项目的 markdown/ 目录。覆盖 10 个数据源：postgresweekly.com/issues（自动解析最新 issue…

    8.6k GitHub stars~1.2k tokensUpdated 2 days ago
    DatabasesAuto-check passed

  • DB AI GitHub Paper Weekly News

    digoal/blog

    抓取并分析"数据库、AI、GitHub、AI论文"相关的近1周内最新内容，分类汇总，输出图文并茂（含内置 Mermaid 图）的 Markdown 周报到当前项目的 markdown/ 目录。

    8.6k GitHub stars~2.4k tokensUpdated 2 days ago
    DatabasesAuto-check passed

  • Esmfold2

    JimLiu/science-skills

    Biohub ESMFold2 / ESMFold2-Fast all-atom co-folding (Candido et al.

    228 GitHub starsUsed in 4 repos~2.5k tokens
    AI & LLM EngineeringAuto-check passed

  • Qwen Mtp Gguf

    R6410418/Jackrong-llm-finetuning-guide

    Complete agent-ready workflow for Qwen-family MTP or nextn GGUF conversion and release.

    1.7k GitHub stars~1.7k tokensUpdated 3 mo ago
    AI & LLM EngineeringAuto-check passed

  • Hugging Face Tokenizers

    Orchestra-Research/AI-Research-SKILLs

    Shows how to load, train and use fast Hugging Face tokenizers, with BPE, WordPiece and Unigram models, padding, truncation and alignment tracking.

    13k GitHub starsUsed in 6 repos~3.4k tokens
    AI & LLM EngineeringAuto-check passed

  • Hugging Face Vision Trainer

    huggingface/skills

    Official

    Trains and fine-tunes object detection, image classification and SAM or SAM2 segmentation models on Hugging Face Jobs cloud GPUs and saves the results to the Hub.

    11k GitHub starsUsed in 1 repo~7.5k tokens
    AI & LLM EngineeringAuto-check passed

More from uw-syfi/TraceLab

  • Coding Trace Analyze

    uw-syfi/TraceLab

    Run, summarize, plot, validate, and export normalized coding-trace JSONL files.

    142 GitHub stars~2.5k tokensUpdated 1 mo ago
    Auto-check passed

  • Coding Trace Collect

    uw-syfi/TraceLab

    Collect, count, and extract Claude Code and Codex CLI local histories into normalized coding-trace JSONL files.

    142 GitHub stars~1.9k tokensUpdated 1 mo ago
    Auto-check: notes

  • Coding Trace Normalize

    uw-syfi/TraceLab

    Explain and work with the normalized coding-trace JSONL row format produced by extractclauderounds.py, extractcodexrounds.py, and collectllmtraces.py.

    142 GitHub stars~1.6k tokensUpdated 1 mo ago
    Auto-check passed

  • Coding Trace Raw

    uw-syfi/TraceLab

    Read and explain raw Claude Code and Codex CLI session logs in this coding-trace repo.

    142 GitHub stars~1.3k tokensUpdated 1 mo ago
    Auto-check passed

  • Coding Trace Sanitize

    uw-syfi/TraceLab

    Sanitize normalized coding-trace JSONL rows for public sharing.

    142 GitHub stars~1.3k tokensUpdated 1 mo ago
    Auto-check: notes

  • Tracelab Public Release

    uw-syfi/TraceLab

    Publish TraceLab snapshots from the internal repo to the public uw-syfi/TraceLab GitHub repo as clean, mergeable, incremental pull requests from a persistent public mirror branch.

    142 GitHub stars~2.7k tokensUpdated 1 mo ago
    Auto-check passed

Works with

Categories

Questions about Publish Tracelab Huggingface

What does Publish Tracelab Huggingface do?

Prepare, publish, refresh, or validate the TraceLab public dataset on Hugging Face under UW-SyFI/TraceLab. Publish Tracelab Huggingface is an agent skill from uw-syfi/TraceLab. Prepare, publish, refresh, or validate the TraceLab public dataset on Hugging Face under UW-SyFI/TraceLab.

When should I use Publish Tracelab Huggingface?

Publish Tracelab Huggingface fits situations like: hugging Face releases; dataset Card updates; viewer/Data Studio failures; loaddataset compatibility.

How do I install Publish Tracelab Huggingface in Claude Code?

Run `npx skills add uw-syfi/TraceLab --skill publish-tracelab-huggingface -a claude-code`. Or copy the skill folder (.codex/skills/publish-tracelab-huggingface in uw-syfi/TraceLab) into .claude/skills/publish-tracelab-huggingface in your project. Claude Code loads it when a task matches its description.

How do I install Publish Tracelab Huggingface in Codex?

Run `npx skills add uw-syfi/TraceLab --skill publish-tracelab-huggingface -a codex`. Or copy the skill folder (.codex/skills/publish-tracelab-huggingface in uw-syfi/TraceLab) into .agents/skills/publish-tracelab-huggingface in your project. Codex loads it when a task matches its description.

Can I use Publish Tracelab Huggingface in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add uw-syfi/TraceLab --skill publish-tracelab-huggingface -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/publish-tracelab-huggingface, .gemini/skills/publish-tracelab-huggingface, .github/skills/publish-tracelab-huggingface and .opencode/skills/publish-tracelab-huggingface in your project.

What does Publish Tracelab Huggingface need to run?

Going by SKILL.md and its folder, Publish Tracelab Huggingface needs Python for the scripts in its folder, the command-line tools its instructions call (uv, gh, git and hf) and credentials named HF_TOKEN. Our summary lists: Python 3.

Does Publish Tracelab Huggingface access the network?

SKILL.md contains no URLs. Its commands use uv, gh and git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Publish Tracelab Huggingface safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Publish Tracelab Huggingface use?

Publish Tracelab Huggingface is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Publish Tracelab Huggingface use?

About 1.4k tokens (SKILL.md is roughly 5.4k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Publish Tracelab Huggingface?

Skills that share tags, products or a category with Publish Tracelab Huggingface: DB AI GitHub Paper Weekly News (digoal/blog, 8.6k stars), DB AI GitHub Paper Weekly News (digoal/blog, 8.6k stars), Esmfold2 (JimLiu/science-skills, 228 stars) and Qwen Mtp Gguf (R6410418/Jackrong-llm-finetuning-guide, 1.7k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Publish Tracelab Huggingface?

uw-syfi (a GitHub organization) maintains it in uw-syfi/TraceLab, which has 142 GitHub stars. The repository holds 7 skills in this directory. The repository was last updated on August 22, 2026.

Source: uw-syfi/TraceLab on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.