drug-discovery

Drug discovery: ChEMBL search, drug-likeness, interactions.

Quality

62%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./optional-skills/research/drug-discovery/SKILL.md

Quality

Content

71%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 a strong, execution-focused skill: all five workflows are copy-paste API pipelines with example inputs, and the reasoning guidelines add genuine domain value. Weaknesses are duplicated endpoint info in the Quick Reference table, missing error-recovery guidance, and two bundled scripts that are orphaned because SKILL.md never mentions them.

Suggestions

Reference the bundled scripts in the relevant workflows, e.g. 'For batch screening use scripts/ro5_screen.py' and 'For target searches use scripts/chembl_target.py', so they are discoverable.

Add a short error-recovery note (what to do when an API call fails, a target is not found, or rate limits are hit) to close the workflow-clarity gap.

Trim the persona preamble and drop or shrink the Quick Reference table, which restates endpoints already visible in the code.

Dimension	Reasoning	Score
Conciseness	The body is dominated by lean, executable curl/python snippets with almost no explanation of concepts Claude already knows. Minor over-explanation remains: the persona framing ('You are an expert pharmaceutical scientist...') and the Quick Reference table, which duplicates endpoints already shown in the code. This is anchor 4 (efficient with minor trims possible), not anchor 5, because the table and roleplay preamble do not earn their tokens.	4 / 5
Actionability	Every workflow is a concrete, copy-paste-ready curl + python pipeline with example IDs (CHEMBL203, CHEMBL25/aspirin, EGFR) and inline parsing of real API response fields. Minor gaps keep it at anchor 4 rather than 5: snippets rely on $1 positional arguments that break when pasted directly into a shell, and curl failures or empty API responses other than the two handled cases are not covered.	4 / 5
Workflow Clarity	Each of the five workflows has a clear, coherent fetch-parse-report sequence, several snippets validate empty results, and the rate-limit note ('add sleep 1 between batch requests') acts as an operational checkpoint; workflows are read-only so the destructive/batch cap does not apply. It is not a 5 because there are no explicit error-recovery or feedback-loop steps (e.g., what to do when an API returns an error or a target is not found).	4 / 5
Progressive Disclosure	The body is well-sectioned and references/ADMET_REFERENCE.md is clearly signaled one level deep ('See references/ADMET_REFERENCE.md for detailed guidance'), but the actual bundle contains scripts/chembl_target.py and scripts/ro5_screen.py which are never referenced anywhere in SKILL.md, making them undiscoverable. Per the guideline to score against the actual bundle structure, orphaned script files place this at anchor 3 (could be better organized) rather than 4 (minor gaps only).	3 / 5
	Total	15 / 20 Passed

Description

53%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.

The description identifies a distinct niche and names real tools, but it is a fragment list rather than a well-formed capability statement, and it entirely lacks a 'Use when...' trigger clause. Trigger coverage misses common synonyms like ADMET, Lipinski, bioactivity, and adverse events.

Suggestions

Add an explicit trigger clause, e.g. 'Use when researching drug candidates, targets, ADMET properties, drug interactions, or adverse events.'

State concrete actions with verbs and include synonyms: 'Search ChEMBL and PubChem for bioactive compounds, calculate drug-likeness (Lipinski Ro5, Veber), and look up drug interactions and adverse events via OpenFDA.'

Use third-person declarative sentences instead of a colon-separated fragment list so both 'what' and 'when' are unambiguous.

Dimension	Reasoning	Score
Specificity	The description names the domain ('Drug discovery') and three terse action fragments ('ChEMBL search, drug-likeness, interactions'), but the fragments lack verbs and comprehensive coverage — 'drug-likeness' does not state what action is performed (calculate? assess?). This matches anchor 3 (domain plus 1-2 concrete actions, not comprehensive) rather than anchor 4, which expects several specific stated actions.	3 / 5
Completeness	A partial 'what' is present (search ChEMBL, drug-likeness, interactions) but there is no 'Use when...' clause or equivalent trigger guidance, which caps completeness at 3 per the judging guidelines. It is not a 2 because the 'what' is reasonably informative rather than vague.	3 / 5
Trigger Term Quality	'Drug discovery', 'ChEMBL', 'drug-likeness', and 'interactions' are natural phrases a pharma researcher would say, but common variations and synonyms are missing: ADMET, Lipinski/Ro5, bioactivity, adverse events, pharmaceutical, PubChem. This matches anchor 3 (some relevant keywords, missing common variations) — not anchor 4, which requires good coverage with only a few terms missing.	3 / 5
Distinctiveness Conflict Risk	Naming ChEMBL carves out a clear pharma-research niche with distinct triggers, giving minimal conflict risk with unrelated skills. It is not a 5 because 'drug-likeness' and 'interactions' are broad enough to overlap with a generic chemistry or scientific-research skill.	4 / 5
	Total	13 / 20 Passed

Validation

81%

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 — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	13 / 16 Passed

Repository: NousResearch/hermes-agent
Path: optional-skills/research/drug-discovery/SKILL.md
Commit: 68fa7e9

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.