From quant-llm-skills
Use when assessing, ranking, comparing, or rating dilution risk for one or more small-cap stocks. Produces a transparent 0-100 score integrating ATM activity, placement-agent tier, filing recency, cash runway, warrant attachment, and repeat-dilution history, with auditable component breakdown and action thresholds (SEVERE / HIGH / MODERATE / LOW / MINIMAL).
How this skill is triggered — by the user, by Claude, or both
Slash command
/quant-llm-skills:dilution-event-scoringThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
A reproducible 0–100 framework for "how dilutive is this name right now."
A reproducible 0–100 framework for "how dilutive is this name right now." The point is not to be perfectly accurate — it is to be transparent and consistent so you can compare across tickers and across time without drift.
Score components, then combine. Never start from gut, then justify.
Each component contributes independently. Final score is a weighted sum, capped at 100. Every score must be reproducible from the inputs; if you can't write down which signals fired, you don't have a score — you have an opinion.
| Component | Max points | Sourced from |
|---|---|---|
| Active ATM detected | 30 | atm-detection |
| Placement agent tier | 20 | bank-tier-classification |
| Filing recency | 15 | filing dates |
| Cash runway | 15 | last 10-Q "Liquidity" / cash + burn rate |
| Structural toxicity (warrants, ELOC, deep discount) | 10 | 424B / 8-K body |
| Repeat-dilution history (12mo) | 10 | filing-count over 12-month window |
| Total | 100 |
| State | Points |
|---|---|
| Confirmed active ATM (3-signal rule passes; recent 10-Q confirms in-use) | 30 |
| Active ATM authorized but no recent 10-Q confirmation of use | 22 |
| ELOC active (Common Stock Purchase Agreement with single counterparty) | 25 |
| Registered direct in last 30 days, no ATM | 12 |
| Effective shelf only (S-3/F-3 effective, no 424B in 6+ months) | 5 |
| No effective shelf, no recent offering | 0 |
Apply only when an offering or sales agreement is in flight. If no agent is named or none of the signals above fire, this component is 0.
| Tier | Points |
|---|---|
| Tier 4 (specialized small-cap placement, e.g., Wainwright on ATM) | 20 |
| Tier 3 (boutique / sector specialist) | 12 |
| Tier 2 (mid-tier full-service) | 6 |
| Tier 1 (bulge bracket) | 2 |
| No agent involved (e.g., S-1 for shell, no underwriter) | 8 (unknown is risk) |
The market's memory of a dilution event decays. Recent = more pressure.
| Most recent dilution-relevant filing | Points |
|---|---|
| 424B / 8-K item 3.02 in last 7 days | 15 |
| In last 30 days | 10 |
| In last 90 days | 5 |
| > 90 days | 0 |
Estimate from last 10-Q: (cash + equivalents) / quarterly burn.
Use the most conservative interpretation when data is partial.
| Estimated runway | Points |
|---|---|
| < 3 months | 15 |
| 3–6 months | 12 |
| 6–12 months | 7 |
| 12–24 months | 3 |
| > 24 months | 0 |
If runway cannot be estimated (no recent 10-Q, going-concern language without numbers), assign 10 and flag for manual review.
Read the 424B / agreement for these features. Each present = 2 points, capped at 10.
Count distinct dilutive events in the trailing 12 months (424Bs, 8-K 3.02s, S-8 effectivenesses, registered amendments).
| Events in last 12 months | Points |
|---|---|
| 6+ | 10 |
| 3–5 | 7 |
| 2 | 4 |
| 1 | 2 |
| 0 | 0 |
| Score | Label | Action recommendation |
|---|---|---|
| 75–100 | 🚨 SEVERE | Treat as continuous selling pressure. Avoid long. Short candidate on strength. Confirm with atm-detection. |
| 55–74 | ⚠️ HIGH | Material dilution overhang. Reduce size, no overnight on green days. |
| 35–54 | ⚠️ MODERATE | Watch carefully. Tradable with discipline. |
| 15–34 | LOW | Background risk only; structure looks clean. |
| 0–14 | MINIMAL | No active dilution mechanism detected. |
Input: ticker XYZ, today.
| Signal | Score |
|---|---|
| ATM detected: confirmed active, last 10-Q says "$8M sold YTD under ATM" | 30 |
| Agent: H.C. Wainwright on the Sales Agreement (Tier 4) | 20 |
| Most recent 424B5: 5 days ago | 15 |
| Cash runway from last 10-Q: ~4 months | 12 |
| Structural toxicity: warrants attached, going-concern language → 2 + 2 | 4 |
| Repeat dilution: 5 events in 12 months | 7 |
| Total | 88 |
Verdict: SEVERE. Continuous selling overhang; treat any rally as a print opportunity for the issuer.
This is not a price target or a trading signal. It quantifies the
structural dilution risk in the filings; it does not predict
short-term price moves, scoring trade entries, or confirm execution
mechanics. Combine outputs with price/volume context and your own
trade plan. Required upstream skills: atm-detection,
bank-tier-classification, sec-filing-types, lookahead-safety.
Searches MemPalace before answering questions about past work, people, projects, or prior decisions. Returns verbatim stored content instead of guessing from model memory.
Guides Payload CMS config (payload.config.ts), collections, fields, hooks, access control, APIs. Debugs validation errors, security, relationships, queries, transactions, hook behavior.
Implements vector databases with Pinecone, Weaviate, Qdrant, Milvus, pgvector for semantic search, RAG, recommendations, and similarity systems. Optimizes embeddings, indexing, and hybrid search.
npx claudepluginhub jefrnc/quant-llm-skills --plugin quant-llm-skills