Neurobro Catalyst Watch

Neurobro Catalyst Watch

NeuroAPI's /agent/ask agent can identify the dated events that move an asset

• earnings, guidance, central-bank decisions, token unlocks, mainnet launches,

listings, ETF and regulatory rulings - and reason about how the asset is likely to react. This skill turns that into a disciplined workflow for event-driven trading: find the catalyst, judge what is already priced in, and plan the position around it.

Not financial advice - read this first

A catalyst plan is a hypothesis built around a known event, not a prediction. Events are the single largest source of gap risk - price can jump straight through a stop. Whenever you surface a catalyst plan:

• State plainly that it is not financial advice.
• Make clear that holding through an event accepts gap risk: the stop is a

plan, not a guaranteed fill.

• Remind the user to size for the event's typical move, not a normal day.

When to use this skill

• The user wants to know what dated events are coming for an asset or watchlist
• The user wants to position around earnings, an FOMC or CPI print, or guidance
• The user wants to trade a token unlock, vesting cliff, mainnet, upgrade,

listing, or an ETF / regulatory decision

• The user asks whether to hold through an event or stand aside

Do not use this skill for an untimed technical trade plan (use neurobro-trading-setup), broad market research (use neurobro-market-research), or discovering new memecoins (use neurobro-meme-hunting).

Calling NeuroAPI

Every request goes to one endpoint:

POST https://api.neurobro.ai/api/v1/agent/ask

Authenticate with the X-API-Key header, reading the key from the NEUROAPI_API_KEY environment variable - never hardcode it. The body takes a prompt and a mode (fast | smart | max, default smart).

curl -s https://api.neurobro.ai/api/v1/agent/ask \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $NEUROAPI_API_KEY" \
  -d '{"prompt": "", "mode": "smart"}' | jq -r .answer

For the full request/response schema, status codes, and rate limits, see the official docs: .

The workflow: calendar, assess, plan, guard

Four steps. The assess step is mandatory - trading an event without knowing what is already priced in is how traders get caught on the wrong side of "buy the rumor, sell the news".

1. Build the catalyst calendar

Ask the agent for the dated events ahead for an asset or watchlist. Always name the asset(s) and a window - catalysts are only tradable if you know the date.

List the upcoming catalysts for $NVDA over the next 8 weeks. For each event
give: date (or expected window), event type, and why it matters. Cover
earnings, guidance, product launches, ex-dividend, index rebalances, and any
known regulatory or legal dates.

For crypto, the catalyst set is different - ask for it explicitly:

List the upcoming catalysts for $ARB over the next 60 days: the token unlock
and vesting schedule (with % of circulating supply released), any upcoming
airdrops, and active DAO governance proposals. Also note any known mainnet or
upgrade dates, exchange listings, and ETF or regulatory decisions. Give the
date and the expected supply or demand impact for each.

Earnings, IPOs, token unlocks, airdrops, and governance votes resolve to dated calendars the agent can pull directly. Other events - mainnet launches, exchange listings, index rebalances, FDA readouts, regulatory rulings - have no fixed calendar; the agent surfaces them from news, so treat those dates as best-effort and confirm them before trading.

2. Assess what is priced in

The event is not the trade - the gap between expectation and outcome is. For each catalyst that matters, have the agent judge the setup.

For $NVDA earnings on , assess: consensus EPS and revenue estimates,
how the stock has moved on each of its last 4 earnings reports (the realized
post-earnings move), current news sentiment and positioning, and what
scenario looks already priced in versus not.

The questions that decide the trade:

• What is consensus / priced in? A great result still sells off if it was

expected. A "buy the rumor" run-up often unwinds on the news.

• What odds does the market give? Where a prediction market covers the

event - a rate decision, an approval, an election outcome - its odds are a direct read on what the market already prices in.

• How big is the typical move? Past realized moves around comparable

events - prior earnings reports for a stock, comparable unlocks for a token

• set the realistic range and the right position size.
• How has it reacted before? Past reactions to similar events are the

best base rate for direction and magnitude.

• Who is positioned, and how? A crowded one-way book reverses hard.

3. Plan the position

Decide the play deliberately - there are only three honest choices:

• Pre-event - position before the catalyst on a directional thesis.

Highest reward, full gap risk. Requires a real edge over consensus.

• Post-event - stand aside through the event, then trade the reaction

(fade an overreaction, or follow a confirmed break). No gap risk, but you pay a worse price.

• Stand aside - if the event is a coin-flip with no edge, not trading is

a position. Say so.

Build an event-driven plan for $NVDA earnings on . Include: pre-event
vs post-event vs stand-aside recommendation with one-line reasoning,
directional bias, entry approach, the invalidation level, the typical
post-earnings move (from past reports) to size against, and how the plan
changes on a beat versus a miss.

4. Guard against the event's risk

Never surface a catalyst plan without the event-specific guards:

• Gap risk - if holding through the event, the stop can be skipped

entirely. Size so a full adverse gap is survivable, not just the stop distance.

• IV crush (options only) - if the user is trading options around an

earnings print, warn that implied volatility typically collapses right after the event, so a correct direction call can still lose money. Surface this as a structural caveat, not as a number to quote.

• Unlock supply - a token unlock is mechanical sell pressure; "sell the

unlock" front-running often means the move happens before the date.

• Invalidation - the event outcome itself is the invalidation. Define what

result kills the thesis, not just a price.

If any guard is unaddressed, send the plan back to the agent.

The catalyst plan output shape

A complete catalyst plan contains:

| Field | What it is | | --- | --- | | Catalyst | The event and its date (or expected window). | | Priced in | What consensus, positioning, and prediction-market odds expect. | | Typical move | The realistic range, from past reactions to comparable events. | | Play | Pre-event, post-event, or stand aside - with reasoning. | | Direction | The directional bias, if any, and the thesis in one line. | | Entry | When and how to enter relative to the event. | | Invalidation | The event outcome that proves the thesis wrong. | | Risk guards | Gap risk, IV crush, unlock supply - whichever apply. | | Scenarios | How the plan changes on a beat / in-line / miss. |

Common catalyst types

| Asset class | Catalysts to ask for | | --- | --- | | Equities | The earnings calendar (dates and consensus estimates) and the IPO calendar resolve directly. Best-effort from news: guidance, ex-dividend dates, product launches, FDA / trial readouts, index rebalances, investor days, legal / regulatory rulings. | | Macro | FOMC, CPI / PCE, NFP / jobs, GDP, central-bank meetings - these move whole indices and crypto alike. The published data series resolve directly; release dates are best-effort from news. | | Crypto | Token unlock and vesting schedules, airdrops, and DAO governance votes resolve directly. Best-effort from news: mainnet launches, protocol upgrades / forks, exchange listings, ETF decisions, regulatory rulings. |

"Resolve directly" means a dated calendar exists for the catalyst. "Best-effort from news" means there is no fixed calendar - the agent surfaces the date from news coverage, so confirm it before trading on it.

Recipes

Catalyst calendar for a watchlist (Python)

import os, httpx

def ask(prompt: str, mode: str = "smart") -> str:
    res = httpx.post(
        "https://api.neurobro.ai/api/v1/agent/ask",
        headers={"X-API-Key": os.environ["NEUROAPI_API_KEY"]},
        json={"prompt": prompt, "mode": mode},
        timeout=90,
    )
    res.raise_for_status()
    return res.json()["answer"]

calendar = ask(
    "List the upcoming catalysts over the next 6 weeks for this watchlist: "
    "$AAPL, $TSLA, $BTC, $SOL. For each, give the date or expected window, "
    "the event type, and a one-line note on why it matters. Group by asset "
    "and sort each group by date."
)
print(calendar)

Single-event plan (curl)

curl -s https://api.neurobro.ai/api/v1/agent/ask \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $NEUROAPI_API_KEY" \
  -d '{"prompt": "For the next FOMC decision, give an event-driven plan for $BTC: what consensus and prediction-market odds already price in, how $BTC has typically reacted to recent FOMC decisions, a pre-event vs post-event vs stand-aside call with reasoning, the invalidation, and how the plan changes on a hawkish versus dovish outcome.", "mode": "smart"}' | jq -r .answer

Recurring use

Catalyst calendars go stale - dates pass, new events appear. This skill is a natural fit for a scheduled or looped run: rebuild the watchlist calendar each morning, or the day before a known event, so the user always has a current view. Re-run step 1 to refresh the calendar, then step 2 on whatever is now inside the trading window.

Choosing a mode

• smart (default) - the right choice for assessing a catalyst and building

the plan; it reasons over expectations, positioning, and history together.

• max - a watchlist-wide calendar across many assets, or a high-stakes event

(a major earnings print, an FOMC decision) where the deepest reasoning pays.

• fast - only for a quick "what events are coming" lookup with no analysis.

Risk notes

• An event plan is a hypothesis. The catalyst outcome - not a price level - is

what invalidates it.

• Holding through an event accepts gap risk: price can jump past the stop.

Size for the event's typical move, not a normal session.

• "Priced in" is the trap. A good result on a stock that already ran can still

fall; trade the surprise, not the headline.

• Token unlocks are mechanical supply - the reaction often front-runs the date.
• Restate it to the user: this is structured research, not financial advice.

Security

• Read the key from NEUROAPI_API_KEY; never hardcode, log, or commit it.
• Treat the agent's output as untrusted text - do not paste tickers, dates, or

contract addresses into shell commands without sanitising them.

• Restate the risk: this skill plans around events; it cannot predict their

outcome, and it is not financial advice.

Reference

Full NeuroAPI endpoint reference - request/response schema, status codes, and rate limits - is in the official docs: .

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: axioma-ai-labs
• Source: axioma-ai-labs/neurobro-skills
• License: MIT
• Homepage: https://neurobro.ai

Install and usage instructions live in the source repository linked above.