watchlist-monitorlisted

# Watchlist Monitor Read-only. `reference/READ-ONLY-POLICY.md` (hard rule) and `reference/RESEARCH-STANDARDS.md` (data efficiency, news freshness, completeness checklist, disclosure) apply. Steps name capabilities — resolve each against `reference/BROKER-CAPABILITIES.md`. Built to be run often, so it must stay cheap: one batched price call for the whole list, and depth only on what's actually in range. `market-pulse` covers what the *market* did; this covers what the user said they were waiting for. ## Steps 1. **Read the watchlist.** `PORTFOLIO-PLAN.md` — **Watchlist / themes** (name, why interested, trigger), plus **Exclusions**, **Risk limits** and **Deployable capital** (a hit with no cash behind it is a different conversation), and the **Decision log** so a name the user passed on isn't re-pitched as news. Empty watchlist → say so and offer `portfolio-plan-builder`. Never substitute a generic idea list for names the user actually chose. 2. **Resolve each trigger into something testable**: a **price level**, a **valuation level** (P/E, yield, market cap), an **event** ("after Q2 results", "once the lock-in expires"), or a **theme criterion** ("any index ETF under 0.5% TER"). Anything vaguer is a finding — say it can't be checked and propose a concrete rewrite. 3. **Price the list in one call.** Symbol-search capability once for unresolved identifiers, then a single batched LTP call for every price-level name. The core output is **