pytdc

Therapeutics Data Commons (PyTDC) for AI-ready therapeutic ML datasets and benchmarks; use it when you need standardized dataset loading, meaningful splits (e.g., scaffold/cold-start), and consistent evaluation for ADME/Toxicity/DTI/DDI or molecular optimization.

Quality

78%

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

Fix and improve this skill with Tessl

tessl review fix ./scientific-skills/Evidence Insight/pytdc/SKILL.md

Quality

Content

68%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.

The body is highly actionable with a runnable end-to-end example and clean references to real bundle files, but it loses points on conciseness (redundant split explanations) and workflow clarity (no validation checkpoints for batch benchmark operations).

Suggestions

Add a validation/verification checkpoint to the workflow, e.g. after get_split assert train/valid/test sizes and check for label leakage, so batch benchmark operations have an explicit feedback loop.

De-duplicate the split-strategy explanation: keep the brief mention in Key Features and defer the parameter detail to references/utilities.md rather than restating it in Implementation Details section 2.

Trim editorial narration such as 'While schemas vary by task, common conventions include' and the 'replace with your own model' code comment to tighten token efficiency.

Dimension	Reasoning	Score
Conciseness	Mostly efficient with executable code and bullets, but split strategies are explained twice (Key Features and Implementation Details section 2) and inline comments like 'replace with your own model' plus 'While schemas vary by task' narration add padding that could be tightened.	3 / 5
Actionability	The Example Usage block is fully executable and copy-paste ready, covering load, scaffold split, train, Evaluator-based MAE, and Oracle scoring, and Implementation Details supplies concrete import/code patterns for the common cases.	5 / 5
Workflow Clarity	The numbered 1-5 example sequence is clear, but there are no validation/verification checkpoints for these batch data-loading and benchmark operations; per the batch-operation cap, workflow clarity cannot exceed 3 without validation feedback loops.	3 / 5
Progressive Disclosure	Good structure with well-signaled one-level-deep references to real bundle files (references/oracles.md, utilities.md, datasets.md and scripts/*.py, all present), but the inline Implementation Details re-cover splits/schemas that also live in the reference files, a minor organization overlap.	4 / 5
	Total	15 / 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, specific description with explicit what/when structure and a distinct therapeutic-ML niche. The main weakness is the second-person 'use it when you need' phrasing, which both violates the third-person voice guidance and slightly weakens trigger naturalness.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete capabilities ('standardized dataset loading, meaningful splits (e.g., scaffold/cold-start), and consistent evaluation for ADME/Toxicity/DTI/DDI or molecular optimization') approaching comprehensive coverage; reduced from 5 to 4 because the second-person phrasing 'use it when you need' triggers the voice penalty.	4 / 5
Completeness	Explicitly answers both 'what' (AI-ready therapeutic ML datasets/benchmarks, loading, splits, evaluation) and 'when' ('use it when you need standardized dataset loading, meaningful splits ... and consistent evaluation') with concrete trigger phrases, matching the top anchor.	5 / 5
Trigger Term Quality	Strong natural domain keywords ('ADME/Toxicity/DTI/DDI', 'scaffold/cold-start', 'molecular optimization') that a therapeutic-ML user would say, but relies on abbreviations and lacks a few plain-language synonyms, stopping short of the comprehensive-with-synonyms anchor.	4 / 5
Distinctiveness Conflict Risk	The 'Therapeutics Data Commons (PyTDC)' niche with ADME/Toxicity/DTI/DDI triggers carves out a clear, specialized domain with minimal overlap risk against other skills.	5 / 5
	Total	18 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: aipoch/medical-research-skills
Path: scientific-skills/Evidence Insight/pytdc/SKILL.md
Commit: 63c61d3

Reviewed: 1 hour 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.