causal-detective

Challenge causal claims through structured threat assessment, counterfactual reasoning, and CausalPy falsification checks. Use when validating whether a causal effect is real or when the user asks "is this effect real?" or "can I trust this result?"

Quality

96%

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

92%Weight 40%Scale 1-3

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, actionable skill body that maps threats to concrete CausalPy checks and defines a clear output pattern. Its only real weakness is progressive disclosure: the three referenced files do not exist in the bundle and the path is malformed.

Suggestions

Create the referenced files (counterfactual_analysis.md, threat_catalog.md, falsification_tests.md) under a references/ directory, or remove the References section if no detail is meant to be externalized.

Fix the link paths from 'reference/...' to 'references/...' so they match the standard bundle directory and resolve correctly.

Confirm each referenced file holds the depth promised by its title (e.g. threat_catalog.md should enumerate the named threats) so navigation is genuinely one level deep.

Dimension	Reasoning	Score
Conciseness	Lean and efficient — it assumes Claude knows what causality and falsification are and devotes every section to procedure, questions, and checks rather than concept explanation.	3 / 3
Actionability	Provides concrete, executable guidance: a table mapping each alternative explanation to a specific CausalPy check (e.g. cp.checks.PreTreatmentPlaceboCheck) plus a fully specified output pattern with named fields.	3 / 3
Workflow Clarity	A clear six-step investigation workflow is explicitly sequenced; since this is analytical rather than destructive/batch work, the absence of a validate-retry loop does not cap the score.	3 / 3
Progressive Disclosure	The body signals three one-level-deep references, but no references/ bundle directory exists and the links use 'reference/' (singular), so the targets are dangling and navigation breaks.	2 / 3
	Total	11 / 12 Passed

Description

100%Weight 40%Scale 1-3

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, third-person description that states concrete capabilities and pairs them with explicit, natural-language trigger conditions. It cleanly answers both 'what' and 'when' with minimal risk of overlapping other skills.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'structured threat assessment, counterfactual reasoning, and CausalPy falsification checks' — rather than vague language.	3 / 3
Completeness	Clearly answers both what (challenge causal claims via the three named methods) and when ('Use when validating whether a causal effect is real...').	3 / 3
Trigger Term Quality	Includes natural user phrasing the rubric rewards, quoting 'is this effect real?' and 'can I trust this result?' as explicit triggers.	3 / 3
Distinctiveness Conflict Risk	Occupies a clear causal-validation niche anchored on CausalPy with distinct triggers, making conflict with other skills unlikely.	3 / 3
	Total	12 / 12 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: 3 missing	Warning

	Total	15 / 16 Passed

Repository: pymc-labs/CausalPy
Path: causalpy/skills/causal-detective/SKILL.md
Commit: 7e23946

Reviewed: 13 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.