serenity-alpha

Serenity Alpha

Core Principle

Do not ask only whether the news is impressive. Ask whether an already-observable demand change can rewrite a smaller company's financial statements.

Use this skill to convert news into a testable alpha hypothesis:

news -> observed demand change -> revenue/profit transmission -> small-cap elasticity -> validation path

Treat outputs as research hypotheses, not investment advice. Verify current prices, market caps, filings, earnings calls, and news from reliable sources before naming securities or making time-sensitive claims.

Optional SEC Data Assist

For U.S.-listed companies, use SEC filings as the factual base for reported fundamentals and management disclosure when available. edgartools is a good optional helper for this step because it can retrieve company filings, XBRL financial statements, filing text, insider transactions, ownership forms, and recent 8-K disclosures.

If the environment does not already have it, install with pip install edgartools or uv pip install edgartools. The import package is edgar, not edgartools. SEC access requires an identity; set EDGAR_IDENTITY="Name email@example.com" in the environment or call from edgar import set_identity; set_identity("name@example.com") before requests.

Minimal usage pattern:

from edgar import Company
 
company = Company("AAPL")
filings =

Use it to support the analysis, not to replace the framework:

• Pull the latest 10-K, 10-Q, and relevant 8-K filings before judging whether a demand change has reached reported numbers.
• Use XBRL financials for historical revenue, gross profit, margin, cash flow, balance sheet, share count, and segment clues.
• Search filing text and MD&A for demand-driver terms, customer concentration, backlog/order commentary, capacity, pricing, supply constraints, and risk-factor changes.
• Check Form 4, 13D/G, or 13F data only as supporting context; do not treat ownership activity as proof of the alpha thesis.
• Keep non-SEC data separate: current price, market cap, valuation multiples, sell-side estimates, TAM, channel checks, and industry pricing usually require other current sources.

When using SEC data, cite the filing form and filing date in the reasoning, and state when the data is stale relative to the news item.

Answer Shape

Start and end with the company that best fits the alpha hypothesis.

• Open with a direct one-sentence call: 最值得优先验证的是：Company / ticker，原因是...
• If several names appear, choose a primary candidate and label others as alternatives or supply-chain comparables.
• If no company is investable yet, say so at the top: 暂时没有足够明确的公司，先观察...
• Close by repeating the same primary company and the single validation condition that would confirm or kill the thesis.

Workflow

1. Separate Talk From Demand

Identify whether the item is merely narrative or a demand inflection. Prioritize evidence that has already happened:

• users are adopting or paying
• enterprises are buying or expanding deployments
• suppliers are shipping, backlogged, or raising prices
• production schedules are tightening
• earnings calls or filings mention the demand
• a niche ecosystem is becoming crowded

If demand is not observable, classify the idea as watchlist-only.

2. Translate Phenomenon Into Financial Lines

Use this sentence structure:

For each candidate, map the mechanism to income-statement or balance-sheet items:

• revenue: units, ASP, customer count, usage, take rate, backlog conversion
• gross margin: mix shift, pricing power, utilization, input costs
• operating leverage: fixed-cost absorption, sales efficiency, R&D leverage
• cash flow: working capital, capex needs, inventory turns, prepayments

3. Prefer Small, Pure, Misclassified Picks

Look for "small shovels," not only obvious mega-cap winners. Rank candidates higher when they have:

• small market cap or low revenue base versus the size of the demand shock
• high business purity to the specific demand vector
• key supply-chain position or scarcity value
• low current investor attention
• stale market label that misses the new role
• financial verification likely within 1-4 quarters

Use this rough screen:

4. Test Market Misclassification

Ask:

• What does the market currently think this company is?
• What could it actually be becoming if the demand change persists?
• Is the new label large enough, durable enough, and close enough to reported numbers to matter?

Misclassification creates alpha only when the relabeling can be validated by numbers, not just story.

5. Build a Verification Chain

Convert the thesis into observable checkpoints. Use the most relevant indicators:

• revenue growth or guidance revisions
• gross-margin improvement or mix disclosure
• backlog, book-to-bill, lead times, or order commentary
• inventory drawdown or channel checks
• utilization, capacity expansion, or capex plans
• ASP changes or pricing commentary
• customer concentration changes
• management mentioning the demand driver unprompted
• competitor/supplier/customer corroboration

Define what would confirm, weaken, or falsify the thesis.

6. Score Alpha Strength

Score each candidate qualitatively or on a 1-5 scale:

Prioritize ideas with high certainty, clear transmission, high purity, high elasticity, and near-term verification.

7. Size By Evidence

Frame position posture as conditional research guidance, not a personalized recommendation:

Mermaid Visualizations

For a full report, include 2-4 Mermaid diagrams when they materially improve comprehension. A short answer or data-limited analysis may use fewer. Do not create a diagram merely to meet a quota.

Prioritize these views:

1. A flowchart from news to observed demand, financial-statement lines, beneficiaries, and the decisive validation condition.
2. A candidate map comparing evidence strength with financial elasticity when at least three candidates have comparable inputs; use quadrantChart only with a nearby comparison table.
3. A 1-4 quarter verification path showing confirm, weaken, and falsify branches.

Apply these rules to every diagram:

• Use fenced mermaid blocks, match the report language, keep node IDs in simple ASCII, and keep labels short.
• Prefer broadly supported flowchart, pie, and stateDiagram syntax. Use xychart-beta, quadrantChart, or timeline only as progressive enhancement and retain the adjacent Markdown table as the fallback.
• Use only evidence and values already stated in the report. Keep names, numbers, units, and probability totals consistent with the surrounding tables; never fill missing data for visual completeness.
• Place each diagram beside the analysis it explains and follow it with a one-sentence takeaway. Keep citations, URLs, dates, and detailed caveats outside the diagram.
• Keep a diagram focused: normally no more than 12 nodes or 8 plotted values. Diagrams supplement rather than replace the written transmission logic, risks, and source trail.

Output Template

When analyzing a news item, use this structure:

Quality Bar

• Anchor the analysis in observable demand, not vibes.
• Prefer named revenue mechanisms over broad themes.
• Distinguish first-order beneficiaries from distant second- or third-order stories.
• Use current sourced data for market caps, financials, prices, and recent filings.
• State uncertainty and falsification conditions clearly.

When references/original-framework.md is consulted, preserve its analytical intent but follow this SKILL.md's current output and visualization rules when the formats differ.