choosing-causalpy-methods

Choose the appropriate CausalPy experiment class from a causal or impact question, data structure, treatment assignment, and identification assumptions. Use before writing analysis code when the method is not yet settled, including plain-English questions about whether a campaign, policy, or intervention worked.

Quality

91%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

96%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 tight, highly actionable design-intake workflow with clear sequencing and explicit output contracts. Its only weakness is that the progressive-disclosure references it points to are not present as actual bundle files.

Suggestions

Create the referenced bundle files (reference/triggers.md, reference/decision_tree.md, reference/experiment_decision_guide.md, reference/method_capability_matrix.md, reference/not_in_causalpy.md, and the six disambiguation cards under reference/disambiguation/) so the signaled references resolve to real content.

Add a short note in the Routing Workflow section clarifying which disambiguation card to consult first when multiple routes are close, so the agent does not have to guess the ordering among the six cards.

Consider adding one concrete end-to-end worked example (e.g., a 'did the campaign work?' ITS routing) inline to anchor the abstract intake factors to a real case.

Dimension	Reasoning	Score
Conciseness	The body is lean and well-organized as tight enumerations (Required Intake, Output Contracts), assumes Claude's competence without explaining causal-inference basics, and every section earns its place — matching the score-5 'lean and efficient' anchor rather than the score-4 'minor instances of over-explanation' case.	5 / 5
Actionability	As an instruction-only design-intake skill it provides concrete, executable guidance: a six-item intake checklist, an explicit routing rule, and four output contracts each with required fields and a defined next step. Per the scoring notes, absence of code is not penalized when guidance is this actionable, matching the score-5 anchor.	5 / 5
Workflow Clarity	The process is clearly sequenced (Required Intake → Routing Workflow → Output Contracts) with explicit checkpoints ('do not write analysis code until the method route is matched or the user has answered the key ambiguity') and four well-defined outcome states, matching the score-5 anchor; no destructive/batch operation is present that would cap the score.	5 / 5
Progressive Disclosure	The overview is clearly structured with one-level-deep, well-signaled references (reference/triggers.md, decision_tree.md, and disambiguation cards), but the referenced bundle files do not actually exist under references/ in the skill directory, so structure is good yet the bundle is incomplete — sitting between the score-4 and score-5 anchors.	4 / 5
	Total	19 / 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 clear, specific, and well-scoped, explicitly covering both what the skill does and when to use it with natural trigger phrasing. It distinguishes itself cleanly from sibling skills via an explicit hand-off boundary.

Dimension	Reasoning	Score
Specificity	Names the domain (CausalPy experiment class) and several concrete inputs it acts on — 'causal or impact question, data structure, treatment assignment, and identification assumptions' — listing multiple specific actions with only minor gaps in coverage, matching the score-4 anchor better than the comprehensive score-5 example.	4 / 5
Completeness	Explicitly answers both what ('Choose the appropriate CausalPy experiment class from...') and when ('Use before writing analysis code when the method is not yet settled, including plain-English questions...'), with concrete trigger phrases — a clear match for the score-5 anchor and not merely the score-4 'when could be more explicit' case.	5 / 5
Trigger Term Quality	Includes natural phrases users would say such as 'plain-English questions about whether a campaign, policy, or intervention worked' and 'the method is not yet settled', giving good keyword coverage; a few synonyms or extension-specific phrasings are missing, so it sits just below the score-5 comprehensive anchor.	4 / 5
Distinctiveness Conflict Risk	It carves a clear niche (CausalPy method selection / design-intake) with distinct, specific triggers and explicit hand-off boundary to 'running-causalpy-experiments', giving minimal conflict risk with other skills — matching the score-5 anchor.	5 / 5
	Total	18 / 20 Passed

Validation

93%

Warnings & errors only

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

Criteria	Description	Result
relative_links	Relative link issues: 13 missing, 6 deeper-than-1-level	Warning

	Total	15 / 16 Passed

Repository: pymc-labs/CausalPy
Path: causalpy/skills/choosing-causalpy-methods/SKILL.md
Commit: f94d7af

Reviewed: 3 days ago

Table of Contents

Discovery Implementation Validation

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.