nanoGPT Training Guide
Orchestra-Research/AI-Research-SKILLs
Walks through nanoGPT, Karpathy's compact GPT implementation: training on Shakespeare, reproducing GPT-2, fine-tuning GPT-2 checkpoints and training on your own text.
A skill your agent uses when designing or reviewing self-distillation workflows for agentic models, including trace collection, teacher or judge feedback, rejection sampling, critique, conversion to…
Install Agentic Self Distillation
$ npx skills add burtenshaw/training-agents --skill agentic-self-distillation -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install burtenshaw/training-agents agentic-self-distillation --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/burtenshaw/training-agents.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/agentic-self-distillation .claude/skills/agentic-self-distillation && rm -rf skills-srcUse ~/.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/
Install the "agentic-self-distillation" agent skill from https://github.com/burtenshaw/training-agents/tree/main/.agents/skills/agentic-self-distillation into .claude/skills/agentic-self-distillation/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "agentic-self-distillation", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/burtenshaw/training-agents/tree/main/.agents/skills/agentic-self-distillationType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add burtenshaw/training-agents --skill agentic-self-distillation -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install burtenshaw/training-agents agentic-self-distillation --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/burtenshaw/training-agents.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.agents/skills/agentic-self-distillation .agents/skills/agentic-self-distillation && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "agentic-self-distillation" agent skill from https://github.com/burtenshaw/training-agents/tree/main/.agents/skills/agentic-self-distillation into .agents/skills/agentic-self-distillation/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "agentic-self-distillation", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add burtenshaw/training-agents --skill agentic-self-distillation -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install burtenshaw/training-agents agentic-self-distillation --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/burtenshaw/training-agents.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.agents/skills/agentic-self-distillation .cursor/skills/agentic-self-distillation && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "agentic-self-distillation" agent skill from https://github.com/burtenshaw/training-agents/tree/main/.agents/skills/agentic-self-distillation into .cursor/skills/agentic-self-distillation/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "agentic-self-distillation", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/burtenshaw/training-agents.git --path .agents/skills/agentic-self-distillation--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add burtenshaw/training-agents --skill agentic-self-distillation -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install burtenshaw/training-agents agentic-self-distillation --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/burtenshaw/training-agents.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.agents/skills/agentic-self-distillation .gemini/skills/agentic-self-distillation && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "agentic-self-distillation" agent skill from https://github.com/burtenshaw/training-agents/tree/main/.agents/skills/agentic-self-distillation into .gemini/skills/agentic-self-distillation/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "agentic-self-distillation", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install burtenshaw/training-agents agentic-self-distillationInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add burtenshaw/training-agents --skill agentic-self-distillation -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/burtenshaw/training-agents.git skills-src && mkdir -p .github/skills && cp -r skills-src/.agents/skills/agentic-self-distillation .github/skills/agentic-self-distillation && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "agentic-self-distillation" agent skill from https://github.com/burtenshaw/training-agents/tree/main/.agents/skills/agentic-self-distillation into .github/skills/agentic-self-distillation/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "agentic-self-distillation", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add burtenshaw/training-agents --skill agentic-self-distillation -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install burtenshaw/training-agents agentic-self-distillation --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/burtenshaw/training-agents.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.agents/skills/agentic-self-distillation .opencode/skills/agentic-self-distillation && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "agentic-self-distillation" agent skill from https://github.com/burtenshaw/training-agents/tree/main/.agents/skills/agentic-self-distillation into .opencode/skills/agentic-self-distillation/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "agentic-self-distillation", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
agentic-self-distillation- GitHub stars
- 153
- Token cost
- ~354 tokens
- SKILL.md length
- 147 words
- Files
- 4 (incl. references)
- Skills in repo
- 6
- Repo updated
- First seen
- Licence
- Apache-2.0
At a glance
A skill your agent uses when designing or reviewing self-distillation workflows for agentic models, including trace collection, teacher or judge feedback, rejection sampling, critique, conversion to…
- Works in 6 steps: Define the task source and base policy. → Collect rollouts with full prompts,… → Verify or judge traces before using them… → …
- Reviewing self-distillation workflows for agentic models
- SKILL.md covers Workflow, Guardrails and References
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Agentic Self Distillation is an agent skill from burtenshaw/training-agents. Use when designing or reviewing self-distillation workflows for agentic models, including trace collection, teacher or judge feedback, rejection sampling, critique, conversion to SFT or preference data, iterative TRL training loops, and safeguards against self-reinforcing errors.
Its SKILL.md is about 350 tokens, which your agent loads only when the skill is triggered. The skill folder holds 5 other files, including reference files (for example `agents/openai.yaml`, `references/distillation-loop.md` and `references/trace-schema.md`).
It sits in AI & LLM Engineering, covering Deep learning. The repository describes itself as: A repo on resources for training agents. The licence is Apache-2.0.
When your agent uses it
- Reviewing self-distillation workflows for agentic models
- Including trace collection
- Rejection sampling
- Conversion to SFT
Example prompts
- “/agentic-self-distillation”
Workflow steps
6 steps, taken from the first numbered list in SKILL.md.
- Define the task source and base policy.
- Collect rollouts with full prompts, actions, observations, tool calls,
- Verify or judge traces before using them as training data.
- Convert accepted traces into SFT messages, chosen/rejected preference pairs,
- Train the next model with TRL and evaluate on held-out tasks.
- Compare against the previous model before promoting the new data recipe.
What it can do on your machine
Read from SKILL.md and the folder at commit ec7cc54. 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
No scripts in the folder and no shell commands in SKILL.md.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Agentic Self Distillation loads about 354 tokens when it runs, and up to ~645 if it reads all its reference files. Until then it costs about 77 tokens; SKILL.md has 147 words of instructions outside code blocks.
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); files beside SKILL.md are not scanned.
SKILL.md
The full file from burtenshaw/training-agents at commit ec7cc54, republished under its Apache-2.0 licence (© burtenshaw). 147 words, ~354 tokens.
Download SKILL.mdSave it as
.claude/skills/agentic-self-distillation/SKILL.md (or your agent's skills folder). This skill also uses 3 other files; get the full folder from GitHub.- name
- agentic-self-distillation
- description
- Use when designing or reviewing self-distillation workflows for agentic models, including trace collection, teacher or judge feedback, rejection sampling, critique, conversion to SFT or preference data, iterative TRL training loops, and safeguards against self-reinforcing errors.
Agentic Self-Distillation
Use this skill to turn verified agent behavior into better post-training data.
Workflow
- Define the task source and base policy.
- Collect rollouts with full prompts, actions, observations, tool calls, outputs, and terminal result.
- Verify or judge traces before using them as training data.
- Convert accepted traces into SFT messages, chosen/rejected preference pairs, or prompt-only tasks with reward metadata.
- Train the next model with TRL and evaluate on held-out tasks.
- Compare against the previous model before promoting the new data recipe.
Guardrails
- Do not distill unverified model outputs directly.
- Keep rejected traces; they are useful for DPO or reward modeling.
- Version teacher model, verifier, prompt, and filter.
- Avoid training on eval tasks or hidden benchmark labels.
- Watch for self-confirming errors where the teacher and judge share the same blind spot.
References
references/distillation-loop.md: loop design.references/trace-schema.md: trace fields and conversion targets.
© burtenshaw, 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 3 other files (references) in .agents/skills/agentic-self-distillation of burtenshaw/training-agents.
- SKILL.md
- agents/openai.yaml
- references/distillation-loop.md
- references/trace-schema.md
Open the folder on GitHubat commit ec7cc54
Compare with similar skills
Agentic Self Distillation 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.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Agentic Self Distillation this skillburtenshaw/training-agents | 153 | — | ~354 | Automated safety check: Pass | Apache-2.0 | |
| nanoGPT Training GuideOrchestra-Research/AI-Research-SKILLs | 13k | 2 repos | ~1.7k | Automated safety check: Pass | MIT | |
| Pieter AbbeelK-Dense-AI/mimeo | 282 | — | ~1.5k | Automated safety check: Pass | MIT | |
| Self SupervisedVectorSpaceLab/AREX-Skill | 331 | — | ~3.2k | Automated safety check: Pass | Apache-2.0 | |
| Add Uint Supportpytorch/pytorch | 104k | 2 repos | ~2.3k | Automated safety check: Pass | Custom licence | |
| Segment Anything Model GuideOrchestra-Research/AI-Research-SKILLs | 13k | 8 repos | ~3.3k | Automated safety check: Pass | MIT |
Similar skills
Pieter Abbeel
K-Dense-AI/mimeo
Applies the reasoning of Pieter Abbeel, robotics and reinforcement learning expert, UC Berkeley professor, and co-founder of Covariant.
Self Supervised
VectorSpaceLab/AREX-Skill
A skill your agent uses for PaddleViT's DINO self-supervised vision-transformer pretraining, multi-crop data contracts, teacher/student configuration, single- or multi-GPU launch planning…
Add Uint Support
pytorch/pytorch
Add unsigned integer (uint) type support to PyTorch operators by updating ATDISPATCH macros.
Segment Anything Model Guide
Orchestra-Research/AI-Research-SKILLs
Guide to using Meta's Segment Anything Model for zero-shot image segmentation with point, box or mask prompts, or automatic mask generation.
Add Op
onnx/onnx
Add a new ONNX operator or update an existing operator to a new opset version.
More from burtenshaw/training-agents
Trl Sft
burtenshaw/training-agents
A skill your agent uses when designing, implementing, reviewing, or debugging supervised fine-tuning with TRL SFTTrainer or trl sft, especially for agentic models trained on chat messages…
Trl Post Training
burtenshaw/training-agents
A skill your agent uses when building, reviewing, or editing TRL post-training workflows for agentic applications, including SFT, DPO, GRPO, RLOO, reward modeling, dataset formats, chat templates…
Hugging Face CLI Workflows
burtenshaw/training-agents
A skill your agent uses when working with Hugging Face CLI or Hub workflows for TRL training, including auth, repositories, uploads, downloads, Jobs, buckets, model persistence, dataset checks…
Openenv Agentic Rl
burtenshaw/training-agents
A skill your agent uses when designing, reviewing, or implementing OpenEnv-style environment interfaces for agentic RL with TRL, including reset/step/state contracts, tasksets, Docker or…
Trackio Observability
burtenshaw/training-agents
A skill your agent uses when instrumenting or inspecting TRL training runs with Trackio, run names, metric schemas, dashboards, logs, grep or ripgrep, SFTP, Hugging Face Job logs, remote artifacts…
Related
Categories
Questions about Agentic Self Distillation
What does Agentic Self Distillation do?
A skill your agent uses when designing or reviewing self-distillation workflows for agentic models, including trace collection, teacher or judge feedback, rejection sampling, critique, conversion to…. Agentic Self Distillation is an agent skill from burtenshaw/training-agents. Use when designing or reviewing self-distillation workflows for agentic models, including trace collection, teacher or judge feedback, rejection sampling, critique, conversion to SFT or preference data, iterative TRL training loops, and safeguards against self-reinforcing errors.
When should I use Agentic Self Distillation?
Agentic Self Distillation fits situations like: reviewing self-distillation workflows for agentic models; including trace collection; rejection sampling; conversion to SFT.
How do I install Agentic Self Distillation in Claude Code?
Run `npx skills add burtenshaw/training-agents --skill agentic-self-distillation -a claude-code`. Or copy the skill folder (.agents/skills/agentic-self-distillation in burtenshaw/training-agents) into .claude/skills/agentic-self-distillation in your project. Claude Code loads it when a task matches its description.
How do I install Agentic Self Distillation in Codex?
Run `npx skills add burtenshaw/training-agents --skill agentic-self-distillation -a codex`. Or copy the skill folder (.agents/skills/agentic-self-distillation in burtenshaw/training-agents) into .agents/skills/agentic-self-distillation in your project. Codex loads it when a task matches its description.
Can I use Agentic Self Distillation 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 burtenshaw/training-agents --skill agentic-self-distillation -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/agentic-self-distillation, .gemini/skills/agentic-self-distillation, .github/skills/agentic-self-distillation and .opencode/skills/agentic-self-distillation in your project.
What does Agentic Self Distillation need to run?
SKILL.md names no scripts, command-line tools or credentials: Agentic Self Distillation is instructions for the agent only.
Does Agentic Self Distillation access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Agentic Self Distillation 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. Review the folder before installing.
What licence does Agentic Self Distillation use?
Agentic Self Distillation 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 Agentic Self Distillation use?
About 354 tokens (SKILL.md is roughly 1.4k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 291 tokens, read only when the agent opens those files.
What are the alternatives to Agentic Self Distillation?
Skills that share tags, products or a category with Agentic Self Distillation: nanoGPT Training Guide (Orchestra-Research/AI-Research-SKILLs, 13k stars), Pieter Abbeel (K-Dense-AI/mimeo, 282 stars), Self Supervised (VectorSpaceLab/AREX-Skill, 331 stars) and Add Uint Support (pytorch/pytorch, 104k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Agentic Self Distillation?
burtenshaw (a GitHub user) maintains it in burtenshaw/training-agents, which has 153 GitHub stars. The repository holds 6 skills in this directory. The repository was last updated on September 13, 2026.
Source: burtenshaw/training-agents on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.