Skip to main content
Glama

Capital Signal: ASX announcements

Market screen

market_screen
Read-onlyIdempotent

Market-wide or sector screen on one dimension in a single call: named mining projects, enterprise value, cash and funding, M&A activity, capital raises, substantial-holder moves, insider trading, or company updates (guidance/buybacks/trading updates/distress/consolidations). Suited to questions about many ASX companies at once, a sector or a commodity group ('who has raised money in lithium lately', 'which gold producers upgraded guidance', 'cheapest companies vs cash'). Rows come back ordered, so the first page holds the most significant rows; when nextCursor is not null there are more rows (pass cursor=nextCursor, or fields='compact' for 500 rows a call). Free.

Insider trading is dimension='dealing'. direction='sales' (the default) returns insider sales ranked in dollars, so sizes compare across tickers; option exercises, vesting events, lapses and nil rows are removed and labelled as not insider selling or buying. direction='buys' returns on- and off-market purchases ranked in dollars, then placement/SPP participation labelled as not a market purchase; internal transfers between a director's own entities, equity grants, takeover mechanics and initial/final notices are not returned as buys. Option exercises are commonly reported in filings as insider buying; the dealing screen separates them from genuine purchases and sales.

Each row is drawn from the whole population in scope, so a dimension shows which companies are absent as well as which qualify.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "dimension": "<dimension>"
    },
    "name": "market_screen"
  }
}

Input Schema

NameRequiredDescriptionDefault
tagNoOptional tag to narrow further, e.g. 'Copper'.
roleNopeople: only board tenures in this role (ned = non-executive director).
sortNopeople: takeovers | excess (default) | worst_excess | dilution | busiest | raised | deep_discount | administrations. raises: 'size' (default), 'discount' or 'date'. intercepts: 'gram_metres' (default), 'grade' or 'gram_metres_asc'. cash: 'runway_asc', 'burn' or 'cash'.
limitNoMax rows (default 25, max 100; responses are capped at 30,000 characters).
maxEvNoev: keep companies at or below this EV in DOLLARS.
minEvNoev only: keep companies at or above this EV in DOLLARS.
sinceNoEarliest event date YYYY-MM-DD (required for every dimension except drawdown and people; without it a sweep covers the whole corpus).
stageNoprojects only: one stage, e.g. 'DFS'.
untilNoraises only: latest announcement date YYYY-MM-DD (default: today).
cursorNoNext page: the nextCursor from the previous response. A non-null nextCursor means more rows follow; null is the last page.
fieldsNo'compact': symbol, sector, date, value, class and docKey only (updates rows also carry period, actualOutcome and actualVsMid) - ~8x more rows per response; suited to market-wide sweeps (up to 500 rows per call).
highToNodrawdown: end of the high's window, YYYY-MM-DD (default today).
sectorNoSector, industry group, or tag — or 'all' for the whole market (default 'all').
minDropNodrawdown: minimum fall from the high as a fraction (0.5 = 50%; default 0.3).
outcomeNoupdates+guidance only: keep rows by guidance-vs-actual result. BEAT, IN_RANGE or MISS: rows scored against the reported result. any_closed: every row whose period has ended, scored or not (a row with null actualOutcome is still waiting on, or has no match to, a results filing).
highFromNodrawdown: start of the window the high is taken from, YYYY-MM-DD (default 52 weeks ago).
minGradeNointercepts: minimum grade in the intercept's unit.
commodityNoprojects, raises and cash: substring match on a mining project's commodities, e.g. 'Lithium' (wider than the Lithium tag). intercepts: the assayed commodity, 'gold' or 'Au'.
dimensionYesWhat to sweep. 'cash' rows carry raisedSinceQuarterEndM, raisesSinceQuarterEnd and runwayAfterRaisesQuarters: a quarterly's runway predates any raise after the quarter end, so those fields extend it. 'raises' returns every placement, SPP, entitlement/rights issue, public offer and convertible in the window, one row per raise, ranked by dollar size (sort='discount' for deepest discount first): amount sought and completed, offer price, discount to the last close before the announcement, attaching options, lead manager, status (ANNOUNCED/CLOSED/EXTENDED/COMPLETED) and every docKey. Event rows (raises, dealing, updates, ma, holders) carry priceChangeDay0Pct and priceChange5dPct: the move from the close before the filing. Cleansing prospectuses, option-only offers and announcements re-released by another issuer are left out. A row with sameOfferAs set is a later leg (e.g. retail) of the offer that docKey announced, so the two amounts are one offer and are not additive. For a commodity ('lithium') pass commodity, not sector. 'holders' rows carry holderType (SUPER, FUND_MANAGER, CORPORATE, CUSTODIAN_OR_LENDER, INDIVIDUAL) and holderGroup (one key per institution across its name variants: 'australiansuper', 'state-street', 'first-sentier'); stock-lending notices are left out unless includeLending=true. 'drawdown' returns companies above minMarketCap that are minDrop (e.g. 0.5 = 50%) or more below their highest close in [highFrom, highTo] (default the last 52 weeks), with the high's date and the latest three price-sensitive filings since it: 'miners over $250m 50%+ off their May highs' = sector='Materials', minMarketCap=250000000, minDrop=0.5, highFrom='2026-05-01', highTo='2026-05-31' (since is not needed). 'intercepts' returns drill intercepts read from drilling results (hole, from, width, grade, unit, commodity, isIncluding, isHistorical, previouslyReported, calledHighGrade, gramMetres = g/t x m), ranked by gram-metres: filter commodity ('gold' or 'Au'), minGramMetres, historical (include|exclude|only), previouslyReported (default exclude: restated earlier results are left out), calledHighGrade; sort='gram_metres_asc' with calledHighGrade=true finds 'high-grade' claims on thin intercepts. 'holder_clusters' returns holders that moved in minCompanies+ companies within windowDays, one row per holder with every move and docKey: for a company building stakes use holderType='CORPORATE' and holderDirection='increase'; for a fund rotating, holderGroup or holderType='SUPER'. 'dealing' (or the legacy 'director_trading') returns insider trading: direction=sales (default) is ranked by DOLLAR size with option exercises, vesting, lapses, consideration issues and nil rows already removed; direction=buys returns insider buying. 'cash' answers 'who is raising next': one row per company (its latest quarterly), with effectiveRunwayQuarters (the stated runway, extended by any raise after the quarter end) and quartersOfCashLeftAsOf (the quarter end the estimate dates from - a June quarterly read in October is four months old). Filter maxMarketCap/minMarketCap, maxRunwayQuarters/minRunwayQuarters, excludeRaisedSinceQuarterEnd, commodity; sort='runway_asc' (default when maxRunwayQuarters is set), 'burn' (default) or 'cash'. 'Materials companies under $100m with under 2 quarters of cash that have not raised since' = sector='Materials', maxMarketCap=100000000, maxRunwayQuarters=2, excludeRaisedSinceQuarterEnd=true. A row with no market cap carries marketCapNote (usually: suspended) and is left out by market-cap filters. 'people' ranks directors by their record across every ASX board they have sat on (one row per person, resolved across name variants; since is not needed): sort='takeovers' (most companies taken over, then premium), 'excess'/'worst_excess' (median annualised share-price return vs CS200EW over completed board tenures of a year or more; minTenures, default 2, of those), 'dilution' (median dilution per year), 'busiest' (most current seats; role='ned' for busiest NEDs), 'raised', 'deep_discount' (raises at 20%+ discount), 'administrations'. Rows carry mergeStrength (strong = joined by a shared holding entity or a bio naming the other company; medium = name rarity alone) and the person's tenures with docKeys. Microcap returns are volatile: an excess median over 2 tenures is an anecdote, so minTenures=3+ gives a steadier median. The first page carries the caveats (survivorship: companies taken over or delisted before the corpus was built are absent). A person's full record is in person_profile.
directionNodealing dimension only: 'sales' (default) or 'buys'. Sales are ranked by DOLLAR size, not share count, so they are comparable across tickers.
raiseTypeNoraises only: one raise type.
historicalNointercepts: historical drilling (default include).
holderTypeNoholders and holder_clusters: one kind of holder.
minAmountMNoraises only: keep raises of at least this many A$ millions.
minTenuresNopeople: minimum completed tenures behind a median (excess, worst_excess, dilution; default 2).
updateTypeNoupdates dimension only: narrow to one update type. distress = administrators, receivers, liquidators, covenant breaches, and rating cuts or negative outlooks that leave the company below investment grade; credit_rating = every other rating action, including investment-grade downgrades (BOQ A- to BBB+). Rows carry severity: SEVERE (insolvency, covenant breach), MODERATE (a rating cut or negative outlook) or MINOR.
windowDaysNoholder_clusters: days within which the moves fall (default 30).
currentOnlyNopeople: only people who sit on at least one board today.
holderGroupNoholders and holder_clusters: one institution across its name variants, e.g. 'australiansuper', 'state-street', 'forrestania-resources' (as returned in holderGroup).
guidanceTypeNoupdates+guidance only: e.g. 'UPGRADE'. INITIATED and REAFFIRMED are not upgrades. UPGRADE/DOWNGRADE leave out cost and spending guidance (capex, exploration budgets, AISC), where a higher number is not better; every row carries metricClass (EARNINGS, REVENUE, PRODUCTION, COST, SPENDING, OTHER) and guidanceDriver (ORGANIC, ACQUISITION, DIVESTMENT, ACCOUNTING): an upgrade driven by an acquisition is not the business doing better. Guidance given before its period closed also carries actualOutcome (BEAT, IN_RANGE or MISS), actual, actualVsMid and resultsDocKey once the results are out, which answers 'did they hit their guidance'.
maxMarketCapNocash: keep companies at or below this market cap in DOLLARS (e.g. 100000000).
minCompaniesNoholder_clusters: at least this many companies (default 2).
minMarketCapNodrawdown and cash: minimum market cap in DOLLARS (e.g. 250000000).
minGramMetresNointercepts: minimum gram-metres (g/t x width).
includeLendingNoholders: include stock-lending notices (default false).
includePassiveNoholder_clusters: include custodians/lenders and the index giants (Vanguard, BlackRock, State Street, Dimensional), left out by default.
maxCashAgeDaysNoev only: ignore cash reports older than this many days.
calledHighGradeNointercepts: only intercepts the announcement calls high-grade (true) or not (false).
guidanceQualityNoupdates+guidance only: 'complete' returns only rows with a stated figure on both sides and an unclosed period — the rows that can answer an old-vs-new question.
holderDirectionNoholder_clusters: only increases or only decreases.
maxRunwayQuartersNocash: keep companies with at most this many quarters of cash left (effectiveRunwayQuarters).
minRunwayQuartersNocash: keep companies with at least this many quarters of cash left ('well funded').
previouslyReportedNointercepts: results the filing restates from an earlier release (quarterly summaries, cited prior holes). Default exclude, so each filing ranks on what it newly reports.
includeSubIntervalsNointercepts: include 'including' sub-intervals (default true).
excludeRaisedSinceQuarterEndNocash: leave out companies that have raised since their quarter end (default false).

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

A4.7/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations only declare the tool read-only, idempotent and closed-world; the description adds far more: free to call, 30,000-character response cap, row ordering with most significant rows first, cursor pagination and the fields='compact' 500-row mode. It also discloses substantive data-handling rules the agent could not otherwise predict, such as option exercises/vesting/lapses being stripped from insider sales and labelled, option-only offers and cleansing prospectuses being excluded from raises, and survivorship bias in people rows.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The opening line is front-loaded with purpose, scope and pagination, then the body is organized by dimension — appropriate for 45 parameters. However it is a dense multi-hundred-word block with some redundancy: the insider-dealing paragraph largely restates what the dimension enum description already covers, and the person/people rules could be tightened.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With 45 parameters and no output schema, completeness is the hardest bar, and the description meets it: it enumerates the fields returned for the key dimensions (raises statuses, holders holderType/holderGroup, intercepts gramMetres, people mergeStrength/minTenures caveats), gives worked filter combinations, and flags edge cases like suspended companies carrying marketCapNote. An agent has enough to call this correctly without guessing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline would be 3, but the description adds real semantic layering beyond the schema: which dimension each filter belongs to, per-dimension sort defaults (runway_asc when maxRunwayQuarters is set, burn otherwise), and non-obvious field meaning like 'a row with sameOfferAs is a later leg ... the two amounts are not additive' and why effectiveRunwayQuarters extends a stated runway. It also warns that 'UPGRADE' excludes cost/spending guidance and that an acquisition-driven upgrade is not organic improvement.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and scope ('Market-wide or sector screen on one dimension in a single call') and then enumerates the twelve dimensions so the agent knows exactly what surface it covers. It explicitly contrasts itself with a sibling ('A person's full record is in person_profile'), so it can be told apart from board/person tools without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives concrete question-to-parameter mappings for nearly every dimension ('who has raised money in lithium lately', 'cheapest companies vs cash', 'miners over $250m 50%+ off their May highs'), plus a disambiguation rule ('For a commodity pass commodity, not sector') and a default-override note for sort per dimension. It implies the multi-company vs single-company boundary ('suited to questions about many ASX companies at once') but never states an explicit when-not or names the sibling to use for single-company questions.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources