SKILLEMALL.ai
GitHub ↗ RU Sign in

AB autoresearch-pilot

Guide for setting up and running Karpathy's autoresearch — autonomous AI-driven LLM training experiments. Helps write program.md, interpret results, and optimize configs for different GPU sizes. Use when: (1) setting up autoresearch, (2) writing or improving program.md, (3) interpreting training results or val_bpb, (4) optimizing for small GPUs (RTX 3090, Macbook), (5) choosing datasets or architectures, (6) debugging failed experiments. Homepage: https://clawhub.ai/skills/autoresearch-pilot

ClawHub Agent Skills author: TommoT2 v1.0.0 MIT-0 2 files body ≈ 1 101 tokens Open the sourceclawhub.ai↗ analyzed 4 d ago

As a process B 65/100 · Nearly there — weak spots: result and completion, when it triggers, progress reporting

GeneratorSoftware developmentInfrastructureAI and agentstype and topics are labelled automatically from the skill text
Runs in: OpenClaw Claude / Claude Code ChatGPT / Codex Gemini CLI Cursor GitHub Copilot / VS Code Qwen Code Kimi Code CLI Z.ai (GLM) DeepSeek (Deep Code) Mistral Vibe LangChain (deepagents) Goose OpenCode
JSON
Technical rating
A
94/100
safety, quality, tests
Safety 60%
100
Quality 40%
84
Run on models
none yet
Process rating
B
65/100
Nearly there
Progress reporting w 2
0
When it triggers w 12
20
Result and completion w 14
40
the three weakest of ten parameters · all ten

How to improve

    For the model run — optional
    • Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
    • A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.

    Guard findings · 0

    ✓ No critical or high findings

    Files scanned: 2. Evidence is masked. Grey chips explain why severity was lowered.

    Against the Agent Skills spec

    ✓ No remarks against the Agent Skills spec

    Process rating: all ten parameters 65/100

    • 0Progress reporting. Says nothing while it works
    • 20When it triggers. No condition that starts the skill
    • 40Result and completion. Does not say what the result is
    • 55Failures and branches. 1 branches
    • 60Tools and files. Uses tools (git, python) that frontmatter does not declare
    • 70Inputs and preconditions. Inputs and preconditions are listed
    • 100Steps. 38 steps
    • 100Consistency. Name and required fields are in place
    • 100Execution cost. Instruction body is 1101 tokens
    • 100Running it twice. No mutating operations

    Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.

    Quality signals

    • +5Description has no quoted example phrases that should trigger the skill
    • +4Description does not say when NOT to use the skill (false activations)
    • +3Output format is not stated: the model decides each time
    • +1No license
    • +2Single-language instructions
    • +3Description length 496: enough signal without eating the budget
    • +4Structure: 13 headings
    • +3Step-by-step instructions: 38 items
    • +4Has examples (1 code blocks)

    Quality base 70; lint remarks subtract, signals add up to 100. Result: 84.

    External checks

    ClawHub: suspicious
    The skill bundle is mostly coherent, but one bundled review helper grants broad local authority and can automatically send code diffs to fallback reviewer tools.
    LLM: suspicious (medium) · VirusTotal: · 29 May 2026