CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

alphafold-database-fetch-and-analyze

Retrieve and analyze AlphaFold predicted structures for a protein. Use when the user provides a specific UniProt Accession ID and wants structural confidence metrics (pLDDT), domain boundary analysis, or disorder assessment. Do not use if the user only has a protein name, gene name, or amino acid sequence — ask for a UniProt ID first.

Quality

87%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

82%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 well-structured, highly actionable guide: every script has a copy-paste-ready command with real example inputs, and the output-interpretation section includes mandatory warning checks. Remaining gaps are minor — some redundancy with the frontmatter description, no error-recovery loop for failed fetches, and an unreferenced references/citation.bib in the bundle.

Suggestions

Trim the 'Do NOT use when' section to only the conflict carve-outs not already stated in the frontmatter description (Foldseek, RCSB PDB, custom predictions) and drop filler phrases like 'Read it carefully'.

Add a brief failure-handling step: what to do when fetch_structure.py cannot find an entry (e.g., verify the UniProt ID, suggest searching UniProt) to close the workflow's error-recovery gap.

Either reference references/citation.bib in the body (e.g., a note to cite the AlphaFold DB paper using it) or remove it from the bundle so all bundled files are discoverable.

DimensionReasoningScore

Conciseness

Quotes: "uv run scripts/fetch_structure.py P00520 -o /path/to/output/" and "Handles fragment fallback for very large proteins" — the body is efficient and adds only skill-specific knowledge. It is not a 5 because the "Do NOT use when" list largely duplicates the frontmatter description and phrases like "Read it carefully and synthesize the results for the user" are trimmable padding.

4 / 5

Actionability

Quotes: "uv run scripts/fetch_structure.py P00520 -o /path/to/output/", "uv run scripts/analyze_plddt.py ./data/AF-P00520-F1-metadata.json", "uv run scripts/analyze_pae.py ./data/AF-P00520-F1-predicted_aligned_error_v6.json" — fully executable, copy-paste-ready commands for all three bundled scripts, with concrete example inputs and an explicit output-path rule; all referenced script files exist in the bundle.

5 / 5

Workflow Clarity

Quotes: "Check the script output for any `[!] WARNING` lines" and "Combine the 'pLDDT Conclusion' and the 'PAE Structural Conclusion'" — a clear sequence (prerequisites → fetch → analyze → interpret) with an explicit warning checkpoint. Not a 5 because there is no error-recovery guidance (e.g., what to do if fetch fails or no entry exists), leaving feedback loops implicit.

4 / 5

Progressive Disclosure

The three scripts are each clearly signaled with descriptions and executable examples, and the body is organized into well-scoped sections (Prerequisites, Overview, Core Rules, Utility Scripts, Interpreting the Output). Not a 5 because the bundle's references/citation.bib is never mentioned anywhere in the body, leaving one bundled resource undiscoverable.

4 / 5

Total

17

/

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.

The description is strong: it explicitly states what the skill does and when to use it with concrete domain triggers (UniProt ID, pLDDT, domain boundaries, disorder), and includes explicit do-not-use boundaries. Its only weaknesses are minor — a few missing capability mentions (PAE download) and a few absent natural term variations.

DimensionReasoningScore

Specificity

Quotes: "Retrieve and analyze AlphaFold predicted structures", "structural confidence metrics (pLDDT), domain boundary analysis, or disorder assessment" — several specific concrete actions are named. It falls short of the 5 anchor because it omits capabilities the skill actually has (PAE matrix retrieval, inter-domain flexibility analysis), leaving minor gaps in coverage.

4 / 5

Completeness

Quotes: "Retrieve and analyze AlphaFold predicted structures for a protein. Use when the user provides a specific UniProt Accession ID and wants..." — both 'what' and 'when' are explicitly and clearly answered with concrete trigger phrases, matching the 5 anchor; it even adds negative triggers ("Do not use if the user only has a protein name...").

5 / 5

Trigger Term Quality

Quotes: "UniProt Accession ID", "pLDDT", "domain boundary analysis", "disorder assessment" — good coverage of the natural terms a user in this domain would say. Not a 5 because common variations like "protein structure", "PAE", or file terms (.cif, mmCIF) are absent.

4 / 5

Distinctiveness Conflict Risk

Quotes: "AlphaFold predicted structures", "specific UniProt Accession ID", "Do not use if the user only has a protein name, gene name, or amino acid sequence" — a clear niche with distinct triggers and explicit carve-outs against neighboring tools (Foldseek, RCSB PDB, custom AlphaFold runs), giving minimal conflict risk.

5 / 5

Total

18

/

20

Passed

Validation

93%

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

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

Total

15

/

16

Passed

Repository

google-deepmind/science-skills

Path

skills/alphafold_database_fetch_and_analyze/SKILL.md

Reviewed

Table of Contents

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.