Recent Stock Analysis Checklist

Use this checklist after reading SKILL.md when rebuilding context from only stock names or a short watchlist.

Inputs

Minimum input:

• Stock names or tickers.
• User's time horizon if provided.
• User's position weights, cost basis, or risk tolerance if provided.

Default assumptions when missing:

• Recent price window: last 5 trading days.
• News/catalyst window: last 10 calendar days.
• Context window: last 1-3 months when the move lacks an obvious catalyst.
• Output: research support, not investment advice.

Security Resolution

For each name, create:

Field	Required behavior
display_name	User-facing name.
ticker	Verify from current source when possible.
exchange	HKEX/SSE/SZSE/Nasdaq/NYSE/etc.
listing_type	Common share, ADR, A-share, H-share, ETF/ETP, preferred, bond, unknown.
confidence	high, medium, low, or needs_confirmation.
notes	Ambiguity, dual listing, ticker mismatch, stale mapping risk.

Evidence Pack

Collect and label:

Evidence	Minimum field
Price	latest/last close, date, 1D/5D change, benchmark, sector peer if available.
Volume/liquidity	volume spike or note unavailable.
News	title, URL, source, publication time, claim type.
Filings	exchange/IR/SEC/HKEX/CNINFO/SSE/SZSE URL and date.
Financials	latest revenue, profit, margin, YoY/QoQ when available; otherwise source gap.
Guidance/orders	order/backlog/RPO/capex/capacity, with official/reported status.
Sector lead-lag	US/Japan/Korea/Taiwan/Europe leader move, benchmark, and relationship.
Macro	rates/FX/commodities/policy when directly relevant.

Source Priority

1. Exchange filings, company IR, SEC/HKEX/CNINFO/SSE/SZSE, official releases.
2. Earnings calls, investor presentations, official product pages.
3. Reputable news with named source or documents.
4. Broker/analyst views, channel checks, supply-chain checks.
5. Social media, forums, unsourced screenshots, price action.

Use tiers 1-2 for strong conclusions. Tiers 3-5 can support reported_claim, analyst_view, narrative_shock, or needs_verification.

Classification Rules

Classification	Use when	Do not use when
company_fundamental	Company-specific financials, order, guidance, product, policy, management, or corporate action explains the move.	Only sector beta is visible.
sector_thesis	The move attacks or confirms a sector first-principles chain.	The company has weak exposure to that chain.
overnight_lead_lag	Foreign leaders or non-overlapping market sessions explain local pressure.	Local market already priced the event or no direct link exists.
valuation_or_crowding	Price action reflects stretched expectations, high beta, or crowded unwind without thesis damage.	Primary evidence shows orders/guidance/profit damage.
macro_or_liquidity	Broad risk appetite, rates, FX, credit, or index pressure dominates.	Company/sector event is enough to explain the move.
event_watchlist	Index, ETF/ETP, Stock Connect, rebalance, reconstitution, or product listing event matters.	No flow/corporate-action event exists.
unknown	Evidence is insufficient or contradictory.	The agent is merely reluctant; state what evidence would decide.

Decision Language

Allowed:

• "This is not thesis exit unless..."
• "De-risking is justified if..."
• "This name should be separated from the shared sector bucket."
• "The next verification point is..."
• "The current evidence supports only a narrative shock, not a fundamental break."
• "Apply the sunk-cost rebuy test: if the position were cash now, would the user open the same size before the next risk window? If not, separate the failed or oversized lot from any smaller catalyst lot."
• "Apply close-first timing discipline: on high-disagreement days, morning moves are evidence rather than confirmation; prefer late-session/close confirmation for larger decisions unless a midday setup is independently confirmed and risk-defined."

Avoid:

• "Buy", "sell", "hold", "add now", "cut now".
• Exact position sizing without a user-supplied risk model.
• Price targets unless sourced and clearly labeled as analyst/market estimates.

Output QA

Before final answer, verify:

• Dates are absolute and correct.
• Each ticker/name mapping is explicit.
• Current/recent data was checked for live-market questions.
• Price action is separated from fundamental evidence.
• Company-specific source gaps are visible.
• Sector thesis and individual stock thesis are not conflated.
• A low/static PE is not treated as proof of cheapness; route earnings quality, forward EPS, historical PE, and parent/subsidiary questions to stock-fundamentals-valuation-analyzer.
• The answer contains scenarios and invalidation conditions.
• The answer says it is not investment advice.