dedupe-against-corpus

$npx mdskill add lyndonkl/claude/dedupe-against-corpus

Prevents duplicate seeds by fingerprinting and clustering content.

  • Blocks exact matches and links near-duplicates to existing seeds.
  • Integrates with corpus storage and topic tagging systems.
  • Decides outcomes using SHA256 hashes and Jaccard similarity scores.
  • Returns SKIPPED for duplicates or creates related seed links.

SKILL.md

.github/skills/dedupe-against-corpusView on GitHub ↗
---
name: dedupe-against-corpus
description: Checks a candidate seed against the existing substacker corpus (seeds, drafts, published) for exact duplicates (sha256 fingerprint) and near-duplicates (title Jaccard, first-200-word Jaccard, shared topic cluster). Exact match exits as SKIPPED. Near-match links via related_seeds rather than creating a duplicate. Use after topic tagging and density scoring, before writing the seed. Trigger keywords: dedupe, duplicate, already thought, near-match, related seed, fingerprint.
---

# Dedupe Against Corpus

## Table of Contents

- [Three tiers of match](#three-tiers-of-match)
- [Workflow](#workflow)
- [Worked example](#worked-example)
- [Guardrails](#guardrails)

**Related skills:** Called by `ingest-inbox-item` step 4. Queried ad-hoc by `search-corpus`. Backlinks into matched seeds (the one place this skill writes outside new seeds).

## Three tiers of match

1. **Tier 1 — Exact fingerprint**: sha256 of body matches an existing seed's fingerprint. → `SKIPPED`.
2. **Tier 2 — Title similarity**: normalized Jaccard on title tokens ≥ 0.7. → `LINK` candidate.
3. **Tier 3 — Content similarity**: for seeds sharing ≥2 topic tags, first-200-word Jaccard ≥ 0.5. → `LINK` candidate.

Tier 2 and 3 candidates are unioned (up to 3 total LINK targets).

## Workflow

```
Dedupe one candidate seed:
- [ ] Step 1: Grep all corpus/**/*.md frontmatter for fingerprint match
- [ ] Step 2: If exact match, return SKIPPED
- [ ] Step 3: Normalized title Jaccard against all existing seeds
- [ ] Step 4: For seeds sharing ≥2 topic tags, first-200-word Jaccard
- [ ] Step 5: Union tier-2 and tier-3 candidates, cap at 3
- [ ] Step 6: If any LINK candidates, return LINK with related_seeds list; else CREATE
- [ ] Step 7: For LINK, Edit matched seeds' related_seeds to add this candidate's id
```

### Step 7: Backlink safely

The matched seeds get their `related_seeds` field extended (append, never replace) ONLY IF:
- `manual_edits: false` on the matched seed, OR
- The edit is strictly additive (appending an id to a list is always safe).

Do not touch any other field on the matched seed.

## Worked example

**Candidate**: new seed about "dropout as ensemble" with topics `[regularization, ensembling, dropout]`, first 200 words describing thinned-network averaging.

**Existing corpus**:
- `2026-03-11-l2-as-gaussian-prior` — topics `[regularization, bayesian]`. Title Jaccard = 0.1 (different). Shared tags: 1. Skip content-similarity check.
- `2026-02-08-bagging-in-deep-nets` — topics `[regularization, ensembling]`. Shared tags: 2. First-200-word Jaccard: 0.51 → LINK candidate.

**Output**: `{action: LINK, related_seeds: [2026-02-08-bagging-in-deep-nets]}`.

Side effect: `corpus/seeds/2026-02-08-bagging-in-deep-nets.md` gets its `related_seeds` appended with the new seed's id.

## Guardrails

1. Never merge, rewrite, or delete. Only link. Merging is the writer's call.
2. Never alter an existing seed's body. Only its `links.related_seeds` field.
3. Never cross-link into `corpus/dead/` — dead ideas stay dead.
4. Fingerprint is over the body only (not frontmatter). Retagging must not change fingerprint.
5. The 3-link cap prevents unbounded link chains.
6. Never propose LINK across `status: published` to `status: seed` — published posts are immutable except for typo fixes.

## Quick reference

- Returns one of: `SKIPPED`, `LINK`, `CREATE`.
- LINK can carry up to 3 related-seed ids.
- Side effect: backlinks into matched seeds (additive).

More from lyndonkl/claude

SkillDescription
abstraction-concrete-examplesBuilds 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.
academic-letter-architectGuides 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.
adr-architectureDocuments 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.
adverse-selection-priorProduces a Bayesian prior probability that an offered transaction is +EV for the recipient, given that the counterparty chose to propose it. Applies Akerlof market-for-lemons logic -- if they offered it, they believe it is +EV for them, so the prior that it is +EV for us is materially below 50%. Reusable across trade evaluation, waiver drops (another team dropping a player is also adverse selection), job-offer analysis, M&A, and any "someone offered me this" situation. Use when you receive an unsolicited trade/offer/proposal, analyzing incoming trade prior, evaluating why a counterparty proposed a deal, or when user mentions adverse selection, market for lemons, why did they offer this, incoming trade prior, they proposed it, Bayesian adjustment on received offer.
alignment-values-north-starCreates actionable alignment frameworks that give teams a shared North Star (direction), values (guardrails), and decision tenets (behavioral standards). Enables autonomous decision-making while maintaining organizational coherence. Use when starting new teams, scaling organizations, defining culture, establishing product vision, resolving misalignment, creating strategic clarity, or when user mentions North Star, team values, mission, principles, guardrails, decision framework, or cultural alignment.
analogy-weight-checkFor every analogy in a substacker draft, verifies it carries mechanical weight — the analogy does real work explaining the mechanism, not merely decorates it. Cross-references analogy-catalog.md for novelty (is this analogy reused from a prior post?) and domain fit (biology > organizational > sports preferred; physics/military disfavored). Use whenever an analogy appears in the draft. Trigger keywords: analogy weight, decorative, mechanical weight, reused analogy, catalog check, metaphor check.
answer-uncomfortable-questionTakes one strategic question about substacker ("should we launch paid?", "is this section dead?", "are we writing for the wrong audience?") and produces the mandatory evidence + reasoning + downside triad plus a recommendation. Used 3 times per Growth Strategist review. Trigger keywords: uncomfortable question, strategic question, evidence reasoning downside, triad.
attribute-performanceFor each substacker post that materially over- or under-performs the rolling baseline (|z| ≥ 1.0), produces a plain-English attribution paragraph with calibrated confidence (high / medium / low / unexplained). Considers subject-line effect, topic zeitgeist, external share, day-of-week, length effect, and audience-notes signals. Labels unexplained outliers explicitly rather than fabricating a story. Use after compute-baseline when outlier posts exist. Trigger keywords: attribution, why did this post work, outlier explanation, performance analysis.
auction-first-price-shadingComputes the optimal shaded bid for a first-price sealed-bid auction given a true private value, an estimate of the number of competing bidders N, and a value-distribution assumption. Implements the `(N-1)/N` equilibrium shading rule for uniform private values, adjusts for log-normal or empirical value distributions, layers a risk-aversion adjustment, and caps output against the bidder's remaining budget. Domain-neutral auction theory reusable across fantasy sports (baseball FAAB, NBA/NHL waiver auctions), prediction-market limit sizing, sealed procurement bids, and any blind-bid context. Use when user mentions "first-price auction bid", "sealed bid shading", "(N-1)/N", "FAAB bid amount", "auction shading", "optimal bid first-price", "bid for sealed-bid", "blind bid sizing", or when downstream logic needs a principled shade factor rather than an ad-hoc heuristic.
auction-winners-curse-haircutApplies a Bayesian haircut to a bid valuation for common-value auctions where winning is itself evidence the bidder over-estimated. Takes a raw valuation, a value-type classification (common_value / private_value / mixed), the number of informed bidders N, and a signal-dispersion estimate, and returns an adjusted valuation. Domain-neutral and reusable across fantasy FAAB, prediction markets, M&A bids, ad-auction budgets, and any generic bidding context. Use when user mentions "winner's curse", "common value auction", "valuation haircut", "adverse valuation", "Bayesian bid adjustment", or "over-paying in auction".