CtrlK
BlogDocsLog inGet started
Tessl Logo

tooluniverse-vaccine-design

Computational vaccine candidate design: peptide/subunit vaccines via MHC-I/MHC-II epitope prediction (IEDB), population HLA coverage optimization, B-cell epitope identification, and cross-strain conservation analysis. Use for vaccine epitope prediction, HLA allele coverage, multi-epitope construct design, and immunogenicity assessment. Combines predicted MHC binding with experimentally validated IEDB epitopes for higher-confidence designs.

67

Quality

80%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./plugins/tooluniverse/skills/tooluniverse-vaccine-design/SKILL.md

The canonical home for this skill is tooluniverse-vaccine-design in mims-harvard/ToolUniverse

SKILL.md
Quality
Evals
Security

Quality

Content

67%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 delivers a well-sequenced, genuinely actionable pipeline with real tool parameters and a working bundled script, and honest limitations. Its main costs are a padded immunology-explainer opening with repeated caveats, and inlined reference tables that keep the overview longer than it needs to be.

Suggestions

Trim the "Reasoning Strategy" section to the non-obvious guidance (LOOK UP DON'T GUESS, evidence grading, binding ≠ immunogenicity stated once) and cut the repeated binding-vs-immunogenicity caveat from two of its three locations.

Move the tool table, HLA supertype lists, and IC50/coverage/conservation threshold tables into a references/ file, keeping SKILL.md as a lean workflow overview with one-level-deep pointers.

Add explicit validation checkpoints per phase (e.g., confirm predicted epitopes against `iedb_search_epitopes` before assembly, re-check coverage after adding epitopes) so the sequence includes feedback loops, not just decision rules.

DimensionReasoningScore

Conciseness

The "Reasoning Strategy" paragraph explains immunology Claude already knows ("MHC-I for CD8+ CTL response, MHC-II for CD4+ helper response", surface-exposed targets are better antibody sites), and the "MHC binding does not equal immunogenicity" caveat is repeated three times across the body. This matches "mostly efficient but includes some unnecessary explanation"; it is not 4 because an entire multi-sentence paragraph plus repetition could be trimmed without losing tool-irreplaceable information.

3 / 5

Actionability

Mostly executable guidance: the IEDB call uses real parameters (taxon ID "eq.NCBITaxon:2697049", method "netmhcpan_el", allele "HLA-A*02:01") and the bundled `scripts/population_coverage.py` commands with expected output are copy-paste ready. It is not 5 because several examples are placeholder templates ("[organism]", "[antigen_aa_sequence]", "[variant_in_epitope]") and Phase 4's EnsemblVEP/PubMed guidance is thin relative to the concrete standard set elsewhere.

4 / 5

Workflow Clarity

Clear Phase 0–5 sequence with decision thresholds at each phase (percentile-rank cutoffs, coverage-target table with corrective actions like "Add more epitopes", conservation tiers with "avoid"/"redesign" guidance). It is not 5 because there are no explicit output-validation steps (e.g., verifying returned epitope data or re-running after a failed/empty prediction) — checkpoints are decision rules rather than validation-with-feedback loops.

4 / 5

Progressive Disclosure

The single bundle file (`scripts/population_coverage.py`) is real, clearly signaled in the body, one level deep, and its bundled-default caveat is well explained — good structure overall. It is not 5 because the ~230-line body inlines reference-style material (the 14-tool table, HLA supertype frequency lists, multiple threshold tables) that could be split into a references file, leaving SKILL.md as a leaner overview.

4 / 5

Total

15

/

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, comprehensive in naming concrete pipeline actions, with an explicit "Use for" trigger clause. The only weakness is slightly thin synonym coverage in the trigger terms (e.g., no "T-cell epitope" or "peptide vaccine").

DimensionReasoningScore

Specificity

Lists multiple concrete actions — "MHC-I/MHC-II epitope prediction (IEDB)", "population HLA coverage optimization", "B-cell epitope identification", "cross-strain conservation analysis", "multi-epitope construct design" — covering the full pipeline. It is not the level-4 anchor because there are no minor gaps: each pipeline stage is named as a specific action.

5 / 5

Completeness

Explicitly answers both: the first sentence states what it does (full enumerated pipeline) and the "Use for..." clause gives four concrete when-triggers. This matches the level-5 anchor with both what and when clearly and explicitly stated; it is not 4 because the when-clause is already specific and trigger-phrase-based rather than improvable.

5 / 5

Trigger Term Quality

"Use for vaccine epitope prediction, HLA allele coverage, multi-epitope construct design" gives good natural-phrase coverage a user would plausibly say. It falls short of 5 because common synonyms like "T-cell epitope", "peptide vaccine", "antigen design", or "population coverage" as a standalone phrase are missing.

4 / 5

Distinctiveness Conflict Risk

"Computational vaccine candidate design" via "epitope prediction" and "HLA coverage" is a clear niche with distinct, domain-specific triggers — minimal overlap risk with generic bioinformatics skills. It is not 4 because the vocabulary (MHC-I/MHC-II, B-cell epitope, multi-epitope construct) is specific enough that wrong-skill triggering is unlikely.

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.