CtrlK
BlogDocsLog inGet started
Tessl Logo

tooluniverse-structural-proteomics

Structural biology plus proteomics integration for drug target validation. Combines PDB experimental structures, AlphaFold predictions, GPCRdb, SAbDab antibody structures, ProteinsPlus binding-site prediction, and BindingDB ligand-affinity data. Use for druggability assessment, binding-site characterization, ligand-pocket analysis, structural-confidence scoring (resolution, pLDDT), and antibody-target interface analysis.

72

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.

A dense, well-structured orchestration skill: exact tool names with parameters, parameter-mistake gotchas, artifact filters, and graded evidence/interpretation thresholds make it immediately actionable. The gaps are minor — no worked example invocations, no failure-path guidance when a lookup returns nothing, and a tool inventory that could be split into a reference file.

Suggestions

Add one or two fully worked example invocations with real values (e.g., `PDBeSIFTS_get_best_structures(uniprot_id="P04637")` followed by `RCSBGraphQL_get_structure_summary(pdb_id=...)`) so the workflow phases are copy-paste executable rather than just named.

Add a short failure-path note per workflow — what to do when a phase returns no structures, no ligands, or low-confidence predictions (e.g., fall back to AlphaFold + Foldseek, or report the gap explicitly) to close the feedback-loop gap.

Consider moving the per-tool Tool Inventory and the Interpretation/Limitations tables into a reference file (e.g., references/tool_inventory.md) and keeping only the most-used tools inline, slimming SKILL.md to an overview.

DimensionReasoningScore

Conciseness

The body is lean and table-driven throughout — tool inventories with parameters ('`PDBeSIFTS_get_best_structures` (uniprot_id)'), a gotchas table ('`alphafold_get_prediction/summary` | `uniprot_id` | `qualifier`'), and threshold tables ('Resolution <2.0A (X-ray) / <3.0A (cryo-EM)') — with no explanations of concepts Claude already knows. Every section adds domain-specific knowledge Claude cannot infer, matching the level-5 anchor.

5 / 5

Actionability

Guidance is highly concrete: exact tool names with parameter signatures, correct-vs-mistake parameter pairs, artifact filter lists ('GOL, EDO, SO4, PEG'), and numeric decision thresholds. However, there are no fully executable example invocations with real values (e.g., a sample `PDBeSIFTS_get_best_structures` call with a real UniProt ID like 'P04637' in context), so it sits at 'mostly executable with minor gaps' rather than copy-paste-ready level 5.

4 / 5

Workflow Clarity

Three workflows are clearly phased (Workflow 1 Phases 0–5, Workflow 3 Phases 1–8) with decision rules ('Resolution <2.5A for drug design. X-ray > Cryo-EM > NMR > AlphaFold') and validation checkpoints such as Workflow 3's 'PDBeValidation quality → binding site well-resolved?'. It falls short of level 5 because there are no explicit error-recovery/feedback loops (what to do when a phase returns nothing, e.g., no co-crystal exists) even though the skill is read-only analysis rather than a destructive or batch operation, so the level-3 cap does not apply.

4 / 5

Progressive Disclosure

The single file is well-organized into scannable sections (LOOK UP DON'T GUESS, Tool Inventory, Workflows, Gotchas, Evidence Grading, Interpretation, Limitations) with no nested references, and no bundle files exist so everything is appropriately self-contained. It is above level 3 (structure is good and nothing is buried), but at ~125 lines the tool inventory and interpretation tables are bulk reference material that could arguably live in a separate reference file, keeping it at 'good structure with minor organization gaps' rather than level 5.

4 / 5

Total

17

/

20

Passed

Description

92%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 description: third-person voice, concrete capabilities, named data sources, and an explicit 'Use for' trigger clause covering five use cases. The only weakness is slightly incomplete natural-language synonym coverage (no 'protein structure', 'cryo-EM', or similar variations a user might say).

DimensionReasoningScore

Specificity

The description enumerates concrete actions — 'druggability assessment, binding-site characterization, ligand-pocket analysis, structural-confidence scoring (resolution, pLDDT), and antibody-target interface analysis' — alongside the specific data sources combined (PDB, AlphaFold, GPCRdb, SAbDab, ProteinsPlus, BindingDB). Multiple specific concrete actions with comprehensive coverage matches the level-5 anchor; there is no generic filler to push it down.

5 / 5

Completeness

It explicitly answers 'what' ('Combines PDB experimental structures, AlphaFold predictions, GPCRdb, SAbDab antibody structures, ProteinsPlus binding-site prediction, and BindingDB ligand-affinity data') and 'when' via an explicit 'Use for' clause listing five concrete use cases. Both are explicit and concrete, matching the level-5 anchor; it is not the level-4 case because the 'when' clause is already specific and enumerated rather than needing more detail.

5 / 5

Trigger Term Quality

Natural domain phrases a user would say — 'drug target validation', 'druggability', 'binding site', 'antibody', 'AlphaFold' — are present, but coverage is not comprehensive: common variations like 'protein structure', 'cryo-EM', 'co-crystal', or file/database extensions a user might mention are missing. Good coverage with a few natural terms missing fits the level-4 anchor; it is above the partial-keyword level 3 but below the comprehensive synonym coverage of level 5.

4 / 5

Distinctiveness Conflict Risk

The niche is sharply defined by named databases and use cases (SAbDab antibody structures, GPCRdb, BindingDB ligand-affinity, structural-confidence scoring), giving it distinct triggers with minimal overlap risk against generic structural-biology or proteomics skills. This matches the level-5 anchor for a clear niche with distinct triggers.

5 / 5

Total

19

/

20

Passed

Validation

100%

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
mims-harvard/ToolUniverse
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.