CtrlK
BlogDocsLog inGet started
Tessl Logo

drug-interaction-checker

Check for drug-drug interactions between multiple medications. Trigger when user asks about medication compatibility, "can I take X with Y", drug interactions, contraindications, or safety of combining pharmaceuticals.

86

1.29x
Quality

81%

Does it follow best practices?

Impact

96%

1.29x

Average score across 3 eval scenarios

SecuritybySnyk

Passed

No known issues

SKILL.md
Quality
Evals
Security

Evaluation results

96%

23%

Medication Safety Report for Anticoagulation Patient

CLI invocation with JSON output and file saving

Criteria
Without context
With context

CLI invocation via script

70%

100%

Multiple drugs passed

100%

100%

JSON format flag used

100%

100%

Output saved to file

50%

50%

Output file exists

100%

100%

JSON structure: drugs_checked

0%

100%

JSON structure: interactions array

100%

100%

JSON structure: severity field

100%

100%

JSON structure: mechanism field

100%

100%

JSON structure: recommendation field

0%

100%

JSON structure: summary counts

25%

100%

Exit code check documented

100%

100%

Without context: $0.7090 · 2m 57s · 38 turns · 37 in / 9,223 out tokens

With context: $0.5161 · 1m 18s · 24 turns · 27 in / 4,366 out tokens

100%

27%

Lipid and Cardiovascular Medication Safety Audit

Drug class warnings and pairwise interaction detection

Criteria
Without context
With context

CLI or module invocation

0%

100%

All drugs in single call

0%

100%

Detects same-class warning

83%

100%

Drug class interaction identified

83%

100%

Severity classification present

87%

100%

Mechanism included

87%

100%

Recommendation included

87%

100%

Pairwise completeness

80%

100%

No external API calls

100%

100%

Output captured to file

100%

100%

Without context: $0.4906 · 2m 27s · 29 turns · 149 in / 7,759 out tokens

With context: $0.5752 · 1m 20s · 25 turns · 29 in / 4,242 out tokens

94%

17%

Geriatric Patient Supplement and Medication Review

Supplement interactions and brand name normalization

Criteria
Without context
With context

Uses local script

50%

100%

Supplements passed as drug inputs

100%

100%

Brand names accepted

100%

25%

Multiple drugs in single invocation

100%

100%

Supplement interaction appears in output

100%

100%

Severity for supplement interaction

100%

100%

Mechanism for supplement interaction

100%

100%

Output format specified

0%

100%

Results saved to file

100%

100%

No re-implementation of database

0%

100%

Special population note

100%

100%

Without context: $0.8646 · 3m 2s · 31 turns · 37 in / 10,790 out tokens

With context: $0.4621 · 1m 20s · 18 turns · 310 in / 4,576 out tokens

Repository
aipoch/medical-research-skills
Evaluated
Agent
Claude Code
Model
Claude Sonnet 4.6

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.