stable-baselines3

Production-ready reinforcement learning algorithms (PPO, SAC, DQN, TD3, DDPG, A2C) with scikit-learn-like API. Use for standard RL experiments, quick prototyping, and well-documented algorithm implementations. Best for single-agent RL with Gymnasium environments. For high-performance parallel training, multi-agent systems, or custom vectorized environments, use pufferlib instead.

Quality

86%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

86%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, highly actionable skill body that leans on executable examples and cleanly delegates depth to verified reference and script files. Minor verbosity in introductory prose and the absence of an explicit error-recovery loop are the only weak spots.

Suggestions

Remove or trim the 'Purpose'/'Overview' sentences that define concepts Claude already knows (e.g., what vectorized environments or callbacks are) to improve token efficiency.

Add a validate->fix->retry feedback loop in the project workflow (e.g., after check_env() fails, fix the environment and re-run before proceeding) to reach top workflow clarity.

Add the missing `from stable_baselines3.common.vec_env import SubprocVecEnv` import in the vectorized-environments quick-setup snippet so it is fully copy-paste runnable.

Dimension	Reasoning	Score
Conciseness	Mostly efficient with concrete code and tight bullet lists, but a few 'Purpose'/'Overview' sentences restate concepts Claude already knows (what vectorized environments or callbacks are); trimmable but not padded.	4 / 5
Actionability	Numerous copy-paste-ready, executable code blocks cover the common cases (training, custom envs, vectorized envs, callbacks, evaluation, schedules, HER); only a trivial missing import in one vec-env snippet keeps it from being flawless.	5 / 5
Workflow Clarity	The 'Starting a New RL Project' section gives a clear 8-step sequence with an explicit check_env() validation checkpoint, but lacks a validate->fix->retry feedback loop for error recovery.	4 / 5
Progressive Disclosure	A clear overview body signals one-level-deep references to real files (references/.md, scripts/.py) with a consolidated Resources section, and all referenced paths exist in the bundle.	5 / 5
	Total	18 / 20 Passed

Description

87%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, third-person description that clearly conveys both capability and trigger conditions while proactively disambiguating from a competing skill. Minor keyword gaps keep trigger quality just below maximum.

Dimension	Reasoning	Score
Specificity	Names the domain plus six concrete algorithms (PPO, SAC, DQN, TD3, DDPG, A2C) and the scikit-learn-like API style, giving several specific capabilities with only minor gaps in describing discrete actions.	4 / 5
Completeness	Explicitly states what it does (production-ready RL algorithms with a scikit-learn-like API) and when to use it ('Use for standard RL experiments, quick prototyping', 'Best for single-agent RL with Gymnasium environments'), with concrete trigger phrases.	5 / 5
Trigger Term Quality	Strong, natural keyword coverage including 'reinforcement learning', 'RL', algorithm names, and 'Gymnasium environments'; a few common phrasings like 'train agents' are absent, keeping it just below comprehensive.	4 / 5
Distinctiveness Conflict Risk	Clear niche (single-agent RL with Gymnasium) anchored by named algorithms, plus explicit boundary guidance redirecting parallel training, multi-agent, and custom vectorized env cases to pufferlib, minimizing conflict risk.	5 / 5
	Total	18 / 20 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: K-Dense-AI/scientific-agent-skills
Path: skills/stable-baselines3/SKILL.md
Commit: b2a92ba

Reviewed: 3 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.