CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

experiment

Designing A/B tests: hypothesis docs, sample size, feature flags, significance analysis, CUPED, SRM detection, switchback experiments. Use when hypothesis validation is needed.

Quality

81%

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

73%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, expert-level skill body with an exemplary phased workflow and explicit validation checkpoints. Its weaknesses are redundancy between the metadata comment, Principles, Core Contract, and Never sections, and heavy dependence on reference files that are missing from the evaluated bundle.

Suggestions

Remove the CAPABILITIES_SUMMARY HTML comment and consolidate the repeated peeking/HARKing/pre-registration rules so each appears in exactly one section, cutting the Never list's long explanatory clauses.

Ship the referenced reference/*.md and _common/*.md files in the bundle (or inline the minimal templates and sample-size formula) so the recipe dispatch and Reference Map point at files that actually exist.

Add a brief inline fallback for each recipe (one-line method summary or formula) so the core flow remains executable even when reference files are unavailable.

DimensionReasoningScore

Conciseness

Mostly efficient, expert-level rules with real numbers, but the ~35-line CAPABILITIES_SUMMARY HTML comment duplicates body content, and the peeking/HARKing/pre-registration rules are restated across Principles, Core Contract, and Never with long explanatory clauses.

3 / 5

Actionability

Highly concrete instruction-only guidance with specific thresholds (p < 0.001 SRM, >=80% power, 7-14 day duration, 1/5/25/50/100% ramp) and named methods (mSPRT, Benjamini-Hochberg, delta-method), but the executable artifacts (templates, sample-size calculator) are delegated to reference files absent from this bundle.

4 / 5

Workflow Clarity

The HYPOTHESIZE -> DESIGN -> EXECUTE -> ANALYZE table gives per-phase required actions, key rules, and read paths, with ANALYZE explicitly sequencing 'SRM check -> statistical analysis -> confidence intervals -> recommendations', a halt-and-investigate feedback loop, and freeze/kill-switch checkpoints.

5 / 5

Progressive Disclosure

The Reference Map provides well-signaled, one-level-deep, condition-based pointers across 14 files with recipe-to-reference dispatch, but none of the referenced reference/ or _common/ files exist in the bundle, so the clearly signaled navigation cannot actually be followed.

4 / 5

Total

16

/

20

Passed

Description

80%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 that packs comprehensive, concrete capabilities into one sentence with an explicit use-when clause. Its main weaknesses are the thin, generic trigger clause and missing common synonyms (split test, power analysis, multivariate).

DimensionReasoningScore

Specificity

Enumerates seven concrete capabilities ('hypothesis docs, sample size, feature flags, significance analysis, CUPED, SRM detection, switchback experiments'), giving comprehensive coverage of the experimentation domain with no meaningful gaps.

5 / 5

Completeness

Clearly answers 'what' via the capability list and includes an explicit 'Use when hypothesis validation is needed' clause, but the 'when' is narrow and generic rather than specifying concrete trigger scenarios.

4 / 5

Trigger Term Quality

Includes natural terms users would say ('A/B tests', 'sample size', 'significance'), but common synonyms and variations like 'split test', 'experiment results', 'power analysis', or 'multivariate' are missing, and the when-clause adds no additional trigger phrases.

4 / 5

Distinctiveness Conflict Risk

The capability list establishes a clear experimentation niche with distinct triggers, though the generic 'hypothesis validation' trigger could fire on general validation questions unrelated to A/B testing, creating minor overlap risk.

4 / 5

Total

17

/

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

simota/agent-skills

Path

experiment/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.