codeclaw

A workflow for exporting Claude Code and Codex conversations as structured training data and uploading them to Hugging Face, a platform for sharing machine-learning datasets and models.

In plain words
What is it for?
Preparing exports, reviewing personal information and secrets, confirming them, and managing the resulting dataset.
Why use it?
It guides the export through configuration, privacy checks, confirmation, and upload instead of treating conversation data as ready to publish.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/ychampion/codeclaw/docs
Any agent
npx skills add ychampion/codeclaw --skill docs
Clone the repo
git clone --depth 1 https://github.com/ychampion/codeclaw

Made for: Claude Code, Codex.

Per session 54 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,675 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00054 $0.01675
Opus 5 $0.00027 $0.00838
Sonnet 5 $0.00011 $0.00335
Haiku 4.5 $0.00005 $0.00168

Measured 2d ago against content hash 3374ffc25a61, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

codeclaw scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

docs/SKILL.md · 115 lines

How it starts

The opening of the file, as written. The whole thing — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.

codeclaw Skill

THE RULE

Every codeclaw command outputs next_steps. FOLLOW THEM.

Do not memorize the flow. Do not skip steps. Do not improvise. Run the command -> read the output -> follow next_steps. That's it.

The CLI tracks your stage (1-4: auth -> configure -> review -> done). codeclaw export (push) is gated -- you must run codeclaw confirm first or it will refuse.

Getting Started

Run codeclaw status (or codeclaw prep for full details) and follow the next_steps.

Output Format

  • codeclaw prep, codeclaw config, codeclaw status, and codeclaw confirm output pure JSON
  • codeclaw export outputs human-readable text followed by ---CODECLAW_JSON--- and a JSON block
  • Always parse the JSON and act on next_steps

Key fields:

  • stage / stage_number / total_stages -- where you are
  • next_steps -- follow these in order
  • next_command -- the single most important command to run next (null if user input needed first)

PII Audit (Stage 3)

After codeclaw export --no-push, follow the next_steps in the JSON output. The flow is:

  1. Ask the user their full name -- then grep the export for it
  2. Run the pii_commands from the JSON output and review results with the user
  3. Ask the user what else to look for -- company names, client names, private URLs, other people's names, custom domains
  4. Deep manual scan -- sample ~20 sessions (beginning, middle, end) and look for anything sensitive the regex missed
  5. Fix and re-export if anything found: codeclaw config --redact "string" then codeclaw export --no-push
  6. Run codeclaw confirm with text attestations -- pass --full-name, --attest-full-name, --attest-sensitive, and --attest-manual-scan. It runs PII scan, verifies attestations, shows project breakdown, and unlocks pushing.
  7. Push only after explicit user confirmation: codeclaw export --publish-attestation "User explicitly approved publishing to Hugging Face."

Read the full file on GitHub · 115 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 115 lines · 54 tokens per session scan A 3374ffc25a61

Subscribe to this mod's changes

codeclaw is a skill published in the GitHub repository ychampion/codeclaw (11 stars, last pushed 6mo ago), licensed MIT. It adds 54 tokens to every session and 1,675 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

trulens-dataset-curation

Create and curate evaluation datasets with ground truth for TruLens.

truera/trulens · 19 tokens

portaljs-add-dataset

Add a dataset (CSV, TSV, JSON, or GeoJSON) to an existing PortalJS portal. Appends an entry to datasets.json so the catalog and showcase render it automatically; routes the data by source (local file vs remote URL) — R2 via Git LFS by default, remote URLs by passthrough. Use when registering a new dataset in a…

datopian/portaljs · 84 tokens

digital-health-clinical-asr-build

Stage 2 of the Clinical ASR Flywheel. Use when curating clinical terms, tagging IPA, and synthesizing a NeMo manifest. NOT for scoring (use /digital-health-clinical-asr-eval).

NVIDIA/skills · 54 tokens

tao-convert-dataset-format

Run tao-daft convert to convert NVIDIA TAO DAFT datasets between supported formats. Do not use for non-DAFT data. Use when the user asks to convert a DAFT dataset, change DAFT format, change a TAO dataset format, or run tao-daft convert.

NVIDIA-TAO/tao-skill-bank · 71 tokens

tao-validate-dataset-format

Run tao-daft validate to check NVIDIA TAO DAFT datasets for structure, schema, and cross-reference errors. Do not use for non-DAFT formats. Use when the user asks to validate a DAFT dataset, check DAFT schema, validate a TAO dataset format, or run tao-daft validate.

NVIDIA-TAO/tao-skill-bank · 77 tokens

opencaselaw-maintenance

Use for OpenCaseLaw autonomous maintenance loops, production health checks, scraper/completeness triage, safe deploy decisions, and Codex automation work in this repo. Do not use for ordinary feature work unless the task is about operational safety, corpus completeness, or maintenance automation.

jonashertner/opencaselaw · 62 tokens