Help us improve
Share bugs, ideas, or general feedback.
From skills-for-humanity
Decomposes complex claims into atomic sub-claims, classifies them, and identifies load-bearing parts. Useful for verifying large claims or uncovering hidden assumptions.
npx claudepluginhub human-avatar/skills-for-humanityHow this skill is triggered — by the user, by Claude, or both
Slash command
/skills-for-humanity:s4h-investigation-claim-decompositionThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Complex claims are bundles. "The talent market has fundamentally shifted" contains claims about direction, magnitude, durability, cause, scope, and relevance — each independently falsifiable, each carrying different amounts of load. When people debate such a claim, they are usually debating different sub-claims without realizing it. Decomposition makes the hidden architecture visible: what is b...
Entry point for the investigation toolkit. Routes to the right skill for tracing sources, decomposing claims, auditing evidence, generating counter-hypotheses, or triangulating across independent sources.
Adversarially tests claims by seeking counterevidence, identifying logical fallacies, and demanding empirical proof using Wheat tools and web searches.
Analyzes claims by mapping arguments, auditing evidence quality, detecting logical fallacies and biases, and issuing verdicts. For evaluating research or technical arguments.
Share bugs, ideas, or general feedback.
Complex claims are bundles. "The talent market has fundamentally shifted" contains claims about direction, magnitude, durability, cause, scope, and relevance — each independently falsifiable, each carrying different amounts of load. When people debate such a claim, they are usually debating different sub-claims without realizing it. Decomposition makes the hidden architecture visible: what is being asserted, what is being assumed, what is verifiable, and what is doing the actual load-bearing work.
Step 1: State the Master Claim Write out the claim in its current form. If it was stated imprecisely, note that — but trace it faithfully. Do not pre-interpret; decompose what was actually asserted.
Framing check: Confirm the specific claim before continuing. State what you've identified — the actual claim being decomposed and its apparent domain (factual, causal, normative, etc.) — in one sentence, then use AskUserQuestion:
Step 2: Decompose into Atomic Sub-Claims Break the master claim into the smallest independently assessable statements. Each sub-claim should be:
Common decomposition dimensions:
Aim for completeness: a claim with 3 surface sub-claims usually has 6-10 once decomposed carefully.
Step 3: Classify Each Sub-Claim For each sub-claim, assign a classification:
| Classification | Meaning |
|---|---|
| Verified | Well-established, supported by strong independent evidence |
| Checkable | Testable in principle; evidence exists or could be gathered |
| Contested | Active disagreement in the relevant domain; competing evidence |
| Uncheckable | Unfalsifiable by design (definitional, philosophical, or unprovable) |
| Assumed | Treated as true without being argued; load-bearing but unexamined |
Step 4: Identify the Load-Bearing Sub-Claims
Before narrowing: Show the complete list of classified sub-claims to the user first. Use AskUserQuestion:
Not all sub-claims carry equal weight. Determine:
These are the critical nodes: the sub-claims where investigation effort should concentrate.
Step 5: Map the Dependency Structure Do some sub-claims depend on others being true first? Build a simple dependency map: "Sub-claim C requires sub-claim A and B to hold." This reveals whether the whole claim stands or falls on a single contested premise.
Step 6: Recommend Verification Priorities Given the load map and classifications, what should be verified first? Rank by:
Before proceeding, use the AskUserQuestion tool. State your interpretation of the situation in 1–2 sentences — what is being analyzed and what the core question is — then ask:
Proceed based on their selection. If the user reframes, incorporate the correction before running any analysis.
[The claim as stated]
| # | Sub-Claim | Type | Classification | Notes |
|---|---|---|---|---|
| 1 | [sub-claim] | Factual / Causal / Trend / Comparative / Definitional / Normative / Scope | Verified / Checkable / Contested / Uncheckable / Assumed | [Brief note] |
| 2 | ... |
Highest load:
Critical assumptions:
Use investigation-evidence-audit once you've identified the load-bearing sub-claims and want to evaluate the evidence for those specific parts. Use investigation-source-trace when you want to know who first made the master claim and how it has changed. Claim decomposition is specifically about the internal architecture of a claim — not whether its evidence is good, but what it is actually asserting.
After delivering this output, use AskUserQuestion to offer the next move:
/s4h-investigation-source-trace — Trace each component claim to its source/s4h-investigation-evidence-audit — Audit evidence for each component claim/s4h-logic-check — Validate the inference structure between claims