ml-autopilot
Skill0 starsChanged 4 months agoApache-2.0View on GitHub
Arm autonomous mode for hands-off or budgeted ML work. Use when the user asks for unattended training runs, gives a time budget, or wants iteration to continue without supervision.
In this repository5 other files it gives its agents.
- dataset-auditSkill
- hf-doctorSkill
- hf-jobsSkill
- ml-internSkill
- trackioSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 20
NEVER stop working while budget remains.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
!"${CLAUDE_PLUGIN_ROOT}/scripts/autopilot_start.py" $ARGUMENTS
Autopilot is armed: a stop guard now blocks this session from going idle while budget remains. The guard reconciles against real job state before every block and disarms itself when the deadline passes.
Rules while autopilot is armed:
NEVER stop working while budget remains. Nobody is watching; they expect the entire budget used productively. Do not ask "should I continue?", there is nobody to answer.
Your workflow is a loop, not a checklist. Once you have a working result, keep iterating:
- Research the approach (delegate to hf-researcher, read docs, find examples).
- Implement (write code, set up training per /hf-jobs).
- Train and evaluate.
- Push the model to the Hub, always.
- Improve: tune hyperparameters, try different data, adjust the recipe, try a different approach.
- Go to 1.
HYPERPARAMETER TUNING: do not tune by hand one value at a time. Write one sweep script that launches a grid and evaluates each run automatically. One well-designed sweep beats ten manual experiments.
Out of ideas: go back to the literature. Crawl citation graphs deeper, read methodology sections of papers you have not read, extract new datasets and tricks, combine recipes across papers, re-read the task and the training logs. There is always a paper you have not read, and it probably has a better dataset.
Budget your time: reserve the final 10 minutes for evaluation and model saving. The task is not done until the required output exists on the Hub and you have evaluated it.
Disarm early by deleting the autopilot state: rerun this skill with 0.01 hours to let the deadline lapse.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

