Classify Claim

Assigns each extracted claim to one of five buckets — simplified-correct, simplified-boundary, wrong, contested, overclaim — with low/medium/high confidence and one-sentence rationale. Classification happens before primary-source verification (which confirms, not invents). Use for every claim from claim-extractor. Trigger keywords — classify, bucket, simplified vs wrong, claim type, technical classification.

200 stars 25 forksOther Plugin marketplace

Install

Claude Code plugin
/plugin marketplace add lyndonkl/claude

The repo is a Claude Code plugin marketplace. Run /plugin afterwards to pick the plugin that includes this skill.

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add lyndonkl/claude --skill classify-claim
Manual install
git clone --depth 1 https://github.com/lyndonkl/claude.git /tmp/claude
mkdir -p ~/.claude/skills
cp -r /tmp/claude/skills/classify-claim ~/.claude/skills/classify-claim

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Classify Claim

## Five buckets

- **simplified-correct**: strips detail; underlying claim still holds. Keep.
- **simplified-boundary**: holds in common case, breaks in edge case. Fold break into post.
- **wrong**: flat factual error. Fix.
- **contested**: field actively debating. Hedge.
- **overclaim**: true narrowly, asserted broadly. Scope.

## Examples

- "Softmax turns logits into a probability distribution" → **simplified-correct** (omits calibration caveats but the core claim holds).
- "Attention is O(n²) in memory" → **simplified-boundary** (true naive; FlashAttention is O(N)).
- "Softmax produces calibrated probabilities" → **wrong** (distribution ≠ calibrated probability).
- "LLMs exhibit emergent reasoning with scale" → **contested** (Schaeffer et al. 2023 actively disputes).
- "RAG beats fine-tuning for domain knowledge" → **overclaim** (true for tail factoid recall; false for reasoning).

## Workflow

```
Per claim:
- [ ] Step 1: Match claim to nearest bucket by definition
- [ ] Step 2: Assign confidence: low (pattern-matching, need source), medium (field knowledge suggests), high (sure)
- [ ] Step 3: Write one-sentence rationale
- [ ] Step 4: If low confidence, defer final classification until cross-reference-claim confirms
```

## Guardrails

1. Never classify `wrong` without a post-classification primary-source check (that's `cross-reference-claim`).
Read the full skill on GitHub

Details

Skill name
classify-claim
Path
skills/classify-claim/SKILL.md
Repo updated
1 September 2026

FAQ

What does the Classify Claim skill do?

Assigns each extracted claim to one of five buckets — simplified-correct, simplified-boundary, wrong, contested, overclaim — with low/medium/high confidence and one-sentence rationale. Classification happens before primary-source verification (which confirms, not invents). Use for every claim from claim-extractor. Trigger keywords — classify, bucket, simplified vs wrong, claim type, technical classification.

How do I install the Classify Claim skill?

Claude Code plugin: /plugin marketplace add lyndonkl/claude. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Classify Claim popular?

It lives in lyndonkl/claude, which has 200 GitHub stars and 25 forks.

Related skills

Abstraction Concrete Examples

200

lyndonkl/claude

Builds structured abstraction ladders that translate high-level principles into concrete, actionable examples across 3-5 levels. Bridges communication gaps, reveals hidden assumptions, and tests whether abstract ideas work in practice. Use when explaining concepts at different expertise levels, moving between abstract principles and concrete implementation, identifying edge cases by testing ideas against scenarios, designing layered documentation, decomposing complex problems into actionable steps, or bridging strategy-execution gaps.

Testing & QA25

Academic Letter Architect

200

lyndonkl/claude

Guides the creation of evidence-based academic recommendation letters, reference letters, and award nominations that combine concrete examples, meaningful comparisons, and genuine enthusiasm. Use when writing recommendation letters for students, postdocs, or colleagues, or when user mentions recommendation letter, reference, nomination, letter of support, endorsement, or needs help with strong advocacy and comparative statements.

Research & Science25

Adr Architecture

200

lyndonkl/claude

Documents significant architectural and technical decisions with full context, alternatives considered, trade-offs analyzed, and consequences understood. Creates a decision trail that helps teams understand why decisions were made. Use when choosing between technology options, making infrastructure decisions, establishing standards, migrating systems, or when user mentions ADR, architecture decision, technical decision record, or decision documentation.

Software Development25

Closer Critique

200

lyndonkl/claude

Evaluates the final paragraph of a substacker draft for compression and closing form — bolded maxim, forward-looking question, or compressed mechanism statement. For series posts (frontmatter series — {slug}), verifies the running scoreboard (P&L, Brier, W-L) is present and updated. Use on every draft. Blocks publication of series posts missing the scoreboard. Trigger keywords — closer, closing, last paragraph, bolded maxim, scoreboard, CTA, wrap up, conclusion.

Other25

Category Trend Analyzer

200

lyndonkl/claude

For each spending category, computes current-period spend, 6-month rolling average, year-over-year delta, and budget variance, then flags categories that are outliers (>1.5x rolling average or >130% of budget). Produces a ranked list of categories that grew, shrank, or stayed flat, plus the top transactions driving each outlier. Use for monthly spending reviews, identifying lifestyle creep, evaluating budget adherence, or when user mentions category trends, spending changes, budget variance, or outlier spend.

Other25

Consul

213

TerminalSkills/skills

HashiCorp Consul for service discovery, health checking, and service mesh. Use when the user needs to register services, perform health checks, store configuration in the KV store, or set up Connect for secure service-to-service communication.

Other34

Browse the most popular Claude skills, all other skills, or write your own with the skills generator.