← All skills
autoresearch-agent
90 / 100 · Low risk
Fine to try. Some of what it can do is expected: a coding skill usually runs build and test commands.
By ArogyaReddy/alirezarezvani-claude-skills (GitHub) · coding · 39 stars · scanned at commit dc799ac
Autonomous experiment loop that optimizes any file by a measurable metric. Inspired by Karpathy's autoresearch. The agent edits a target file, runs a fixed evaluation, keeps improvements (git commit), discards failures (git reset), and loops indefinitely. Use when: user wants to optimize code speed, reduce bundle/image size, improve test pass rate, optimize prompts, improve content quality (headlines, copy, CTR), or run any measurable improvement loop. Requires: a target file, an evaluation com…
Works with any agent that supports SKILL.md skills.
What this skill can do
Runs commands on your computer
The author does not mention this. Expected for a coding skill.
Found in 47 places.
Show 17 more
- engineering/autoresearch-agent/SKILL.md, line 226
result = subprocess.run(["my-benchmark", "--json"], capture_output=True, text=True) - engineering/autoresearch-agent/SKILL.md, line 237
python scripts/log_results.py --experiment engineering/api-speed - engineering/autoresearch-agent/SKILL.md, line 287
git clone https://github.com/alirezarezvani/claude-skills.git - engineering/autoresearch-agent/SKILL.md, line 293
./scripts/convert.sh --skill autoresearch-agent --tool codex|gemini|cursor|windsurf|openclaw - engineering/autoresearch-agent/SKILL.md, line 298
clawhub install cs-autoresearch-agent - engineering/autoresearch-agent/evaluators/benchmark_size.py, line 26
result = subprocess.run(BUILD_CMD, shell=True, capture_output=True) - engineering/autoresearch-agent/evaluators/benchmark_size.py, line 34
subprocess.run(DOCKER_BUILD_CMD, shell=True, capture_output=True) - engineering/autoresearch-agent/evaluators/benchmark_size.py, line 35
result = subprocess.run( - engineering/autoresearch-agent/evaluators/benchmark_size.py, line 37
shell=True, capture_output=True, text=True - engineering/autoresearch-agent/evaluators/benchmark_speed.py, line 20
subprocess.run(COMMAND, shell=True, capture_output=True, timeout=120) - engineering/autoresearch-agent/evaluators/benchmark_speed.py, line 25
result = subprocess.run(COMMAND, shell=True, capture_output=True, timeout=120) - engineering/autoresearch-agent/evaluators/build_speed.py, line 20
subprocess.run(CLEAN_CMD, shell=True, capture_output=True, timeout=60) - engineering/autoresearch-agent/evaluators/build_speed.py, line 23
result = subprocess.run(BUILD_CMD, shell=True, capture_output=True, timeout=600) - engineering/autoresearch-agent/evaluators/llm_judge_content.py, line 55
result = subprocess.run( - engineering/autoresearch-agent/evaluators/llm_judge_copy.py, line 78
result = subprocess.run( - engineering/autoresearch-agent/evaluators/llm_judge_prompt.py, line 57
gen_result = subprocess.run( - engineering/autoresearch-agent/evaluators/llm_judge_prompt.py, line 76
judge_result = subprocess.run(
and 27 more places
Changes files outside its own folder
This is disclosed by the author in the skill's SKILL.md or README.
Found in 1 place.
Other copies in this repository
ArogyaReddy/alirezarezvani-claude-skills has 1 more folder with a skill named autoresearch-agent. This page shows the lowest-scoring copy.
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'autoresearch-agent' ] || [ -L ~/.agents/skills/'autoresearch-agent' ]; then echo 'A skill folder named autoresearch-agent already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/ArogyaReddy/alirezarezvani-claude-skills' "$d" && git -C "$d" checkout --quiet 'dc799ac68cfdc454fc55edaf3f054651f645f738' -- 'engineering/autoresearch-agent' && mkdir -p ~/.agents/skills && cp -R "$d"/'engineering/autoresearch-agent' ~/.agents/skills/'autoresearch-agent' && rm -rf "$d"; fi
Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'autoresearch-agent' ~/.claude/skills/'autoresearch-agent' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.
Badge for your README
The badge for ArogyaReddy/alirezarezvani-claude-skills shows the score of its lowest-scoring skill.
[](https://zalavijaysinh.github.io/skilltrust/repos/ArogyaReddy/alirezarezvani-claude-skills/)
Score history
| Date | Version | Score |
|---|
| 2026-11-08 | dc799ac | 90 |
| 2026-11-06 | dc799ac | 90 |