Skip to content

CULT+MATH Decision Briefs MCP server

com.cultmath/decision-briefscultmath.comstreamable-httpv1.3.0Scanned

Odds and forecasts on the numbers executives plan against, plus free studies and Decision Briefs.

35 tools: 29 change things and 6 only read.

Tools: 0 delete, 0 irreversible, 29 write, 6 read.
  • Reads private data · 1
  • Reads text others write · 31
  • Can send data out · 2
  • 28 findings

Tools listed without sign-in. Remote posture

With Anlyon

29 tools here change things. Govern them.

  • Anlyon lists the tools and adopts none.
  • Each tool you adopt gets approvals and receipts.
  • Your agent never holds the key.
Govern this server →

A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.

What the CULT+MATH Decision Briefs MCP server can do

The CULT+MATH Decision Briefs MCP server is published by cultmath.com, a domain the MCP Registry verified. It has 35 tools. It can create or change data (29 tools: buy_decision_evidence, run_filing_trend, run_regulatory_trend, run_spend_trend and 25 more), but cannot delete data. 6 tools only read.

1 tool reads private data (get_document); 31 tools bring in text other people can write (buy_decision_evidence, get_decision_evidence, get_document and 28 more); 2 tools can send data out of your system (get_document and run_comment_themes). On its own it has all three parts of a prompt-injection exfiltration path: private data, text others write, and a way out.

The fixed checks matched 28 findings (annotation-conflict). Each one comes with the text it matched.

Its remote endpoint lists its tools without sign-in, and its tools changed once since Anlyon Radar first saw them.

Last scanned Nov 14, 2026, version 1.3.0.

Tools

Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.

35 tools

Changes things 29Claude Code: ask
  • buy_decision_evidenceStarts a Stripe checkout for the full evidence behind one decision: sourced and dated figures, what would flip the answer, and the brief's stated limits.Reads text others writeNo findings

    Why it is in this group

    • The description's first verb is "start": a word for a change.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Starts a Stripe checkout for the full evidence behind one decision: sourced and dated figures, what would flip the answer, and the brief's stated limits. Returns a checkout URL for the user to approve and pay, and a session_id. After the user pays, call get_decision_evidence with both ids. Always show the price and get the user's go-ahead before sending them to pay.

    Annotations the server declares

    • title: Buy one decision
    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • run_ai_visibilityUse when the user asks whether their brand or website shows up in AI answers, or who AI assistants cite instead.Reads text others write1 finding

    Why it is in this group

    • The name "run_ai_visibility" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_ai_visibility" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_ai_visibility" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether their brand or website shows up in AI answers, or who AI assistants cite instead. Asks ChatGPT, Claude, Gemini and Perplexity the buyer's question live with web search on, three times each on the everyday model and once on the flagship model, and returns per platform: in how many askings the domain was cited and at what positions, whether that held every time, and whether the flagship model agrees. Also how many platforms cite it and the domains cited most instead (not the answers themselves). A platform that cites nothing is reported as not measured. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $1.47 per run over x402 (USDC on Base, no account needed), or $1.80 over MPP (a Stripe shared payment token). Fixed price per run, not per answer. Charged in full when at least 3 of 4 platforms are measured (a platform counts when at least 2 of its 3 askings answer with sources); otherwise it refuses and nothing is charged. So a charged run can rest on fewer than all 16 answers; the result reports exactly how many, per platform. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/ai-visibility.json

    Annotations the server declares

    • title: Do AI assistants cite us when buyers ask?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_ai_visibility_quickUse when the user asks whether their brand or website shows up in AI answers and a quick read is enough.Reads text others write1 finding

    Why it is in this group

    • The name "run_ai_visibility_quick" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_ai_visibility_quick" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_ai_visibility_quick" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether their brand or website shows up in AI answers and a quick read is enough. Asks ChatGPT, Claude, Gemini and Perplexity the buyer's question live with web search on, twice each on the everyday model, and returns per platform: in how many askings the domain was cited and at what positions and whether that held both times, plus the domains cited most instead (not the answers themselves). No flagship-model comparison; that is run_ai_visibility. Buy more runs to cover more questions. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.74 per run over x402 (USDC on Base, no account needed), or $1.06 over MPP (a Stripe shared payment token). Fixed price per run, not per answer. Charged in full when at least 3 of 4 platforms are measured (a platform counts when both of its 2 askings answer with sources); otherwise it refuses and nothing is charged. So a charged run can rest on fewer than all 8 answers; the result reports exactly how many, per platform. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/ai-visibility-quick.json

    Annotations the server declares

    • title: Do AI assistants cite us? (quick, under $1)
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_angle_demandUse when the user is choosing a positioning, a message or a campaign angle and asks which benefit or reason to buy matters most to buyers, or which is gaining.Reads text others write1 finding

    Why it is in this group

    • The name "run_angle_demand" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_angle_demand" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_angle_demand" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user is choosing a positioning, a message or a campaign angle and asks which benefit or reason to buy matters most to buyers, or which is gaining. Returns, for each angle searched with the category, its share of searches among the angles named, its year-on-year change and its growth beyond an everyday control basket, with a volume band. Shares and bands only. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.07 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/angle-demand.json

    Annotations the server declares

    • title: Which reason to buy do people search for, and which is growing?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_answer_shareUse when the user asks which brands or websites AI assistants rely on for their category, who they are competing with in AI answers, or whether any source dominates.Reads text others write1 finding

    Why it is in this group

    • The name "run_answer_share" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_answer_share" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_answer_share" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks which brands or websites AI assistants rely on for their category, who they are competing with in AI answers, or whether any source dominates. Asks ChatGPT, Claude, Gemini and Perplexity 2-3 buyer questions live, each twice, and returns the most-cited domains as counts of answers, whether the category is owned, led, contested or fragmented, whether the leaders hold from one asking to the next, and, if given, where your domain stands. Not the answers themselves. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $2.20 per run over x402 (USDC on Base, no account needed), or $2.54 over MPP (a Stripe shared payment token). Fixed price per run, not per answer. Charged in full when at least half the answers cite sources and at least 3 of 4 platforms do; otherwise it refuses and nothing is charged. So a charged run can rest on as few as 12 of 24 answers (fewer with two questions); the result reports exactly how many (controls.measurement). To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/answer-share.json

    Annotations the server declares

    • title: Who owns the answer in my category?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_answer_share_quickUse when the user asks which brands or websites AI assistants rely on for their category and a quick read is enough.Reads text others write1 finding

    Why it is in this group

    • The name "run_answer_share_quick" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_answer_share_quick" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_answer_share_quick" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks which brands or websites AI assistants rely on for their category and a quick read is enough. Asks ChatGPT, Claude, Gemini and Perplexity two buyer questions live, once each, and returns the most-cited domains as counts of answers, whether the category is owned, led, contested or fragmented, and, if given, where your domain stands. Not the answers themselves. No stability check; buy a second run, or use run_answer_share, to see whether the leaders hold. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.74 per run over x402 (USDC on Base, no account needed), or $1.06 over MPP (a Stripe shared payment token). Fixed price per run, not per answer. Charged in full when at least half the answers cite sources and at least 3 of 4 platforms do; otherwise it refuses and nothing is charged. So a charged run can rest on as few as 4 of 8 answers; the result reports exactly how many (controls.measurement). To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/answer-share-quick.json

    Annotations the server declares

    • title: Who owns the answer in my category? (quick, under $1)
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_channel_mixUse when the user is planning go-to-market or channel strategy and asks where buyers look, whether to prioritise retail, review sites, their own site or communities, or who control…Reads text others write1 finding

    Why it is in this group

    • The name "run_channel_mix" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_channel_mix" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_channel_mix" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user is planning go-to-market or channel strategy and asks where buyers look, whether to prioritise retail, review sites, their own site or communities, or who controls the first page for their category. Searches Google live for 1-3 buyer queries and classes every top-10 organic result by domain: retailers and marketplaces, reviews and media, brand sites, forums and social, reference, independent and specialist sites. Returns the share each class holds, who leads each class, and who takes the top three places. Domains and counts only. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/channel-mix.json

    Annotations the server declares

    • title: Where do buyers find options when they search?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_claim_checkUse when the user is choosing a positioning, a tagline or a message and asks whether competitors already say it, whether a claim is crowded or open, or who has moved onto it recent…Reads text others write1 finding

    Why it is in this group

    • The name "run_claim_check" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_claim_check" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_claim_check" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user is choosing a positioning, a tagline or a message and asks whether competitors already say it, whether a claim is crowded or open, or who has moved onto it recently. Reads each competitor's home page today and its Internet Archive copy from a year ago, and returns, per site, whether it leads with the claim, mentions it, or does not, and whether that changed this year, with a verdict: open, mentioned but not owned, taken, or crowded. No page text is returned. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/claim-check.json

    Annotations the server declares

    • title: Is this claim already taken by a competitor?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_comment_themesUse when the user wants to know what viewers raise under a trailer, launch video, ad or review on YouTube, for example whether they talk about price, a feature or a complaint.Reads text others writeCan send data out1 finding

    Why it is in this group

    • The name "run_comment_themes" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe name "run_comment_themes" says "run" and "comment": it writes something other people read.

    Findings

    • Warn

      The server marks "run_comment_themes" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_comment_themes" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user wants to know what viewers raise under a trailer, launch video, ad or review on YouTube, for example whether they talk about price, a feature or a complaint. Returns, for up to eight themes named in advance, the count and share of up to 400 top comments per video (up to five videos) that mention each, with how many comments were read. No comment text. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run.

    Annotations the server declares

    • title: What are people saying under this video?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_cost_squeezeUse when the user asks whether a commodity shock (cocoa, coffee, sugar, wheat, corn, beef, aluminum, oil) has squeezed a public company's margin, whether it passed the cost through…Reads text others write1 finding

    Why it is in this group

    • The name "run_cost_squeeze" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_cost_squeeze" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_cost_squeeze" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether a commodity shock (cocoa, coffee, sugar, wheat, corn, beef, aluminum, oil) has squeezed a public company's margin, whether it passed the cost through, or what a price spike will do to a supplier. Returns the commodity's average price change and the company's gross margin for each fiscal year since 2019, with a verdict: squeezed, held, or widened. Refuses before you are charged if the commodity moved less than 25%. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run.

    Annotations the server declares

    • title: Has this input cost reached the company's margin?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_customer_concentrationUse when the user asks how much of a public company's revenue comes from one retailer or customer (for example Walmart), or is weighing the risk of selling through a dominant buyer…Reads text others write1 finding

    Why it is in this group

    • The name "run_customer_concentration" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_customer_concentration" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_customer_concentration" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks how much of a public company's revenue comes from one retailer or customer (for example Walmart), or is weighing the risk of selling through a dominant buyer. Returns the largest customer's share of revenue for each fiscal year tagged in the company's three latest 10-Ks, the customer's name as tagged, and each filing's citation. Refuses before you are charged if no share is tagged. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run.

    Annotations the server declares

    • title: Is this company dependent on one big customer?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_deal_activityUse when the user asks about M&A activity, deal flow or consolidation in a sector, whether now is a good time to buy or sell a company in it, or who the active acquirers are.Reads text others write1 finding

    Why it is in this group

    • The name "run_deal_activity" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_deal_activity" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_deal_activity" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks about M&A activity, deal flow or consolidation in a sector, whether now is a good time to buy or sell a company in it, or who the active acquirers are. Returns, quarter by quarter, the 8-K merger-agreement documents that mention the sector as a share of all 8-K merger agreements, from SEC EDGAR full-text search, whether that share is rising or falling, an industry-code cross-check, and the most active filers with their latest filing linked. It counts documents, not deals, and gives no deal values or multiples. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/deal-activity.json

    Annotations the server declares

    • title: Is M&A in this sector rising or falling, and who is doing it?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_demand_checkUse when the user asks whether interest in a product, category or trend is really growing, or wants to know if a rise in search is more than noise before investing in it.Reads text others write1 finding

    Why it is in this group

    • The name "run_demand_check" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_demand_check" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_demand_check" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether interest in a product, category or trend is really growing, or wants to know if a rise in search is more than noise before investing in it. Returns the category's year-on-year search change against an everyday basket of ten unrelated searches (its growth beyond the market), a volume band, and whether SEC 10-K filings confirm the direction, the costly-to-fake check. Changes and bands only, never monthly volumes. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.07 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/demand-check.json

    Annotations the server declares

    • title: Is demand for this real, or is search moving for everyone?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_disclosure_checkUse when the user asks whether any company, or which of a named set of competitors, now reports a figure (a franchise share, ARPU, a retention rate, a concentration, a multiple), o…Reads text others write1 finding

    Why it is in this group

    • The name "run_disclosure_check" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_disclosure_check" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_disclosure_check" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether any company, or which of a named set of competitors, now reports a figure (a franchise share, ARPU, a retention rate, a concentration, a multiple), or whether a public comparable discloses the metric a deal model depends on. Returns, from SEC EDGAR full-text search, which filers use the phrase in their filings over the years given, with the first and latest filing linked; for named peers, who does, who does not, and who could not be measured because they filed nothing. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/disclosure-check.json

    Annotations the server declares

    • title: Has anyone started reporting this number?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_fad_checkUse when the user is writing a creative or campaign brief, or weighing a trend, and asks whether a moment, ingredient, format or cultural theme is a lasting shift or a fad that has…Reads text others write1 finding

    Why it is in this group

    • The name "run_fad_check" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_fad_check" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_fad_check" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user is writing a creative or campaign brief, or weighing a trend, and asks whether a moment, ingredient, format or cultural theme is a lasting shift or a fad that has peaked. Returns, for US Google Trends and English Wikipedia interest since 2016, how much of its best 12 months the subject keeps today, with controls, with months since the peak and a verdict: lasting, fading, past its peak, still rising, or signals disagree. Ratios and dates only. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.07 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/fad-check.json

    Annotations the server declares

    • title: Is this moment lasting, or past its peak?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_filing_trendUse when the user asks whether a risk, technology or theme is spreading through public companies, or whether companies now treat it as material.Reads text others write1 finding

    Why it is in this group

    • The name "run_filing_trend" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_filing_trend" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_filing_trend" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether a risk, technology or theme is spreading through public companies, or whether companies now treat it as material. Returns, year by year, the share of U.S. 10-K filers that mention the phrase, from SEC EDGAR full-text search, with a denominator control ("competition") and a null control ("shrinkflation") that must read 0. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/filing-trend.json

    Annotations the server declares

    • title: Is this becoming material to companies?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_launch_liftUse when the user asks whether a line extension, spin-off, sequel, collaboration or new product lifted interest in the parent brand or show, or is weighing a launch on the promise …Reads text others write1 finding

    Why it is in this group

    • The name "run_launch_lift" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_launch_lift" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_launch_lift" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether a line extension, spin-off, sequel, collaboration or new product lifted interest in the parent brand or show, or is weighing a launch on the promise of a halo. Returns, for the launch month, the parent's US search change against the three months before, the best same-window change in every other year since 2019 as a seasonal control, and a verdict against a +20% bar set before the pull: lifted, seasonal, or no lift; optionally English Wikipedia as a second signal. Ratios only. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run.

    Annotations the server declares

    • title: Has this launch lifted the brand it came from?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_peer_benchmarkUse when the user names a public company and asks how its growth, margin, R&D or marketing spend compares with competitors.Reads text others write1 finding

    Why it is in this group

    • The name "run_peer_benchmark" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_peer_benchmark" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_peer_benchmark" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user names a public company and asks how its growth, margin, R&D or marketing spend compares with competitors. Returns the company's revenue growth, gross margin, or R&D or sales-and-marketing as a share of revenue for its latest fiscal year, each peer's, the peer median, the gap and the rank, from SEC EDGAR XBRL as filed; every excluded peer is named with the reason. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/peer-benchmark.json

    Annotations the server declares

    • title: Are we ahead of or behind our peers on this metric?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_premium_checkUse when the user asks whether their price premium is holding, whether customers are trading down, or whether a price rise is hurting demand.Reads text others write1 finding

    Why it is in this group

    • The name "run_premium_check" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_premium_check" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_premium_check" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether their price premium is holding, whether customers are trading down, or whether a price rise is hurting demand. Returns the share of the brand's US Google searches that look for a cheaper way in (price, cheap, discount, coupon, dupe, alternative, sale), against the year before, with a verdict (eroding, holding, strengthening) and each modifier's trend. Shares and bands only. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/premium-check.json

    Annotations the server declares

    • title: Is our premium holding, or are buyers hunting for a cheaper way in?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_regulatory_clockUse when the user asks whether a U.S.Reads text others write1 finding

    Why it is in this group

    • The name "run_regulatory_clock" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_regulatory_clock" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_regulatory_clock" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether a U.S. federal rule on a topic is open for comment, when the deadline is, or when a final rule takes effect. Returns, as of today, the Federal Register documents mentioning the phrase whose comment period is still open (with days left) and the final rules still waiting to take effect, optionally for one agency, each linked. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/regulatory-clock.json

    Annotations the server declares

    • title: Is there still time to respond to a rule on this?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_regulatory_trendUse when the user asks whether rules are coming on a topic, or how much federal rule-making mentions it, optionally for one agency.Reads text others write1 finding

    Why it is in this group

    • The name "run_regulatory_trend" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_regulatory_trend" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_regulatory_trend" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether rules are coming on a topic, or how much federal rule-making mentions it, optionally for one agency. Returns, year by year, the share of U.S. federal final and proposed rules that mention the phrase, from the Federal Register, against every final and proposed rule that year, with a null control that must read 0. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). For deadlines and comment windows, use run_regulatory_clock. Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/regulatory-trend.json

    Annotations the server declares

    • title: Are regulators moving on this?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_revenue_mixUse when the user names a public company and asks how its revenue splits (by platform, revenue type, channel, product or region) or whether its mix is shifting, for example toward …Reads text others write1 finding

    Why it is in this group

    • The name "run_revenue_mix" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_revenue_mix" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_revenue_mix" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user names a public company and asks how its revenue splits (by platform, revenue type, channel, product or region) or whether its mix is shifting, for example toward mobile, live services or digital. Returns every revenue split the company files that adds up to reported revenue across its three latest 10-Ks, with each fiscal year's shares and dollar values and each filing's citation, from SEC EDGAR XBRL as filed. If the company files no split this run can verify, it refuses before you are charged. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run.

    Annotations the server declares

    • title: Where does this company's revenue come from, and how is the mix moving?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_series_forecastUse when the user is setting a plan, budget or inventory for the next year and needs where US retail sales, prices (CPI), employment and wages, industrial production, housing or ve…Reads text others write1 finding

    Why it is in this group

    • The name "run_series_forecast" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_series_forecast" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_series_forecast" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user is setting a plan, budget or inventory for the next year and needs where US retail sales, prices (CPI), employment and wages, industrial production, housing or vehicle sales are heading. Returns a 12-month forecast with an 80% band, next to "same month last year", from a model backtested on the series' own last three years. Refuses before you are charged when last year's same month forecast it better, and says by how much. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run.

    Annotations the server declares

    • title: What will this US series do over the next 12 months?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_spend_trendUse when the user asks whether federal money backs a market, technology or policy, or wants evidence beyond announcements.Reads text others write1 finding

    Why it is in this group

    • The name "run_spend_trend" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_spend_trend" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_spend_trend" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether federal money backs a market, technology or policy, or wants evidence beyond announcements. Returns the share of U.S. federal prime contract obligations matching the keyword, by fiscal year, from USAspending.gov, against all contract obligations. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/spend-trend.json

    Annotations the server declares

    • title: Is the government actually spending on this?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_threshold_oddsUse when the user's plan turns on whether rent, food prices, household energy, wages or retail sales will be above or below a level soon: "what are the chances rent inflation is ab…Reads text others write1 finding

    Why it is in this group

    • The name "run_threshold_odds" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_threshold_odds" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_threshold_odds" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user's plan turns on whether rent, food prices, household energy, wages or retail sales will be above or below a level soon: "what are the chances rent inflation is above 3% in December?", "will grocery store sales grow more than 2%?". The operating numbers a plan is built on, not headline CPI. Returns a probability in percent and in words, the central forecast and its 80% range, and the same question answered by "same month last year" and by the historical base rate, so you can see what the model adds. Up to 3 months past the latest published figure only: further out, the method missed by up to 21 points in testing. Refuses before you are charged when the series fails its benchmark today or the month is out of range. Each sold call is scored when the month is published. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run.

    Annotations the server declares

    • title: What are the odds this US series crosses a line in the next three months?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_trade_concentrationUse when the user asks whether supply of a product has really diversified, how exposed US imports are to one country, or who gained share after tariffs or reshoring.Reads text others write1 finding

    Why it is in this group

    • The name "run_trade_concentration" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_trade_concentration" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_trade_concentration" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether supply of a product has really diversified, how exposed US imports are to one country, or who gained share after tariffs or reshoring. Returns, for two years, the top-three partners' share of US imports of the HS code, the largest single source, a concentration index, and the partners that gained and lost the most share, from UN Comtrade. Refuses before you are charged if the partner rows do not add up to the world total. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Currently switched off: it answers "not available" and charges nothing.

    Annotations the server declares

    • title: Where do US imports of this product come from now?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_triangulationUse when the user needs to know whether a trend is substantive or hype before acting on it.Reads text others write1 finding

    Why it is in this group

    • The name "run_triangulation" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_triangulation" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_triangulation" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user needs to know whether a trend is substantive or hype before acting on it. Returns, year by year, the share of 10-K filers mentioning it (SEC EDGAR), of federal final and proposed rules (Federal Register) and of federal contract obligations (USAspending), each line's direction, and whether the three agree. If any source fails its controls the whole run refuses. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run. A real sample of this run, free: https://cultmath.com/agent/samples/triangulation.json

    Annotations the server declares

    • title: Is this real? Companies, regulators and money checked together
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • run_write_downsUse when the user asks whether a company has impaired a brand or an acquisition, is preparing to sell a brand to a strategic buyer, or wants to know how acquirers have valued past …Reads text others write1 finding

    Why it is in this group

    • The name "run_write_downs" says "run": a word for a change.
    • The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.

    Findings

    • Warn

      The server marks "run_write_downs" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_write_downs" says "run": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Use when the user asks whether a company has impaired a brand or an acquisition, is preparing to sell a brand to a strategic buyer, or wants to know how acquirers have valued past deals. Returns every fiscal year in which the company tagged a goodwill or intangible impairment in its 10-K XBRL, with the amount by line (goodwill, brands and trademarks, other intangibles) and each filing's citation. Refuses before you are charged if none is tagged. Safe to cite: its controls are checked first, it states its limits, and every result carries a run id, a date and a citation line. If a control fails it refuses and you are not charged. This tool charges nothing: it validates the request and returns a payment_url, the price and instructions. Pay by POSTing to payment_url over MPP (HTTP 402, a Stripe shared payment token from your wallet). Price: $0.05 per run over x402 (USDC on Base, no account needed), or $0.50 over MPP (a Stripe shared payment token); charged only when the run comes back sound. To have a sound result considered for publication on cultmath.com, add publish: true and question (what you were asked) to the POST body, and credit_name to be credited; CULT+MATH reviews every submission first, and a published piece earns one free run.

    Annotations the server declares

    • title: Has this company written down what it bought?
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
Only reads 6Show 6 read-only tools
  • get_decision_briefEverything public about one brief: the six decisions it covers with the evidence line for each, pages, edition, price, sample pages, free answers, guarantee, license and payment li…No findings

    Why it is in this group

    • The name "get_decision_brief" says "get": a word for reading.
    Description, as the server wrote it
    Everything public about one brief: the six decisions it covers with the evidence line for each, pages, edition, price, sample pages, free answers, guarantee, license and payment link. The PDF itself is delivered only after payment.

    Annotations the server declares

    • title: Get a Decision Brief
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
  • get_decision_evidenceReturns the full evidence for one decision once it is paid for.Reads text others writeNo findings

    Why it is in this group

    • The name "get_decision_evidence" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Returns the full evidence for one decision once it is paid for. Pass the session_id from buy_decision_evidence, or the Checkout Session id of a full brief, team license, pack or all-access purchase that includes this decision.

    Annotations the server declares

    • title: Get paid evidence
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • get_documentEvery published finding (question and answer) of one CULT+MATH study or answer page, with its date, license, citation line and related brief.Reads private dataReads text others writeCan send data outNo findings

    Why it is in this group

    • The name "get_document" says "get": a word for reading.

    What it can reach

    • private-dataThe name "get_document" says "get" and "document": it reads documents.
    • untrusted-inputThe input takes a URL (input.url), and the tool reads it: it reads web pages, which anyone can write.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.

    Claude Code: ask, because it is tagged outbound.

    Description, as the server wrote it
    Every published finding (question and answer) of one CULT+MATH study or answer page, with its date, license, citation line and related brief.

    Annotations the server declares

    • title: Read a study or answer
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
  • list_decision_briefsList the CULT+MATH Decision Briefs: who each is written for, its headline finding, price, page and payment link.No findings

    Why it is in this group

    • The name "list_decision_briefs" says "list": a word for reading.
    Description, as the server wrote it
    List the CULT+MATH Decision Briefs: who each is written for, its headline finding, price, page and payment link. Also lists bundles and the custom brief.

    Annotations the server declares

    • title: List Decision Briefs
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
  • list_decisionsThe 54 decisions inside the briefs (six per brief), each with who owns it, who contests it, and the public evidence line.No findings

    Why it is in this group

    • The name "list_decisions" says "list": a word for reading.
    Description, as the server wrote it
    The 54 decisions inside the briefs (six per brief), each with who owns it, who contests it, and the public evidence line. Use an id with buy_decision_evidence to buy the full evidence for just that decision.

    Annotations the server declares

    • title: List decisions
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
  • search_cultmathSearch the free CULT+MATH studies and answers (and the briefs) by keywords.No findings

    Why it is in this group

    • The name "search_cultmath" says "search": a word for reading.
    Description, as the server wrote it
    Search the free CULT+MATH studies and answers (and the briefs) by keywords. Returns titles, URLs, dates, the lead finding and the brief that goes deeper.

    Annotations the server declares

    • title: Search CULT+MATH research
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false

This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.

Exfil surface

Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.

62 paths in this server alone from private data, through text someone else wrote, to a way out. Denying 1 private-data tool breaks every one.

Reads private data1
Reads files, mail, messages, databases, secrets, calendars or repositories.
Can send data out2
Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
deny to break every path
{
  "permissions": {
    "deny": [
      "mcp__decision-briefs__get_document"
    ]
  }
}

Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.

Remote posture

What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.

Tools listed without sign-in

Looked at Nov 15, 2026, 13:20 UTC, unchanged since Nov 15, 2026. No credentials were sent and no tool was called. What is looked at.

Reachable
yes
HTTP status of an anonymous initialize
200
tools/list served without credentials
yes
Tools that change things, listed without credentials
29
Protocol version
2025-06-18
WWW-Authenticate on 401
not observed
resource_metadata in WWW-Authenticate
not observed
RFC 9728 protected resource metadata
https://cultmath.com/.well-known/oauth-protected-resource/mcp
Authorization servers
https://cultmath.com
Legacy HTTP+SSE transport
no
Dynamic client registration
yes
Client ID metadata documents
no
  • WarnThe whole server

    The endpoint served tools/list without credentials, and the list has 29 tools that change things (29 write). Whether a call also works without credentials was not tested: tools are never called.

    Matched text

    https://cultmath.com/mcp: tools/list returned 35 tools to an anonymous client, including buy_decision_evidence (write), run_filing_trend (write), run_regulatory_trend (write), run_spend_trend (write), run_triangulation (write) and 24 more

    What this check looks for

Posture changes

  1. Nov 15, 2026
    • Tools that change things, listed without credentials: 18 to 29
    • RFC 9728 protected resource metadata: none to https://cultmath.com/.well-known/oauth-protected-resource/mcp
    • Authorization servers: none to https://cultmath.com
    • Dynamic client registration: not observed to yes
    • Client ID metadata documents: not observed to no

Findings

Matches of fixed checks, with the text they matched.

28 warn
  • The server marks "run_filing_trend" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_filing_trend" says "run": a word for a change.

    What this check looks for
  • The server marks "run_regulatory_trend" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_regulatory_trend" says "run": a word for a change.

    What this check looks for
  • The server marks "run_spend_trend" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_spend_trend" says "run": a word for a change.

    What this check looks for
  • The server marks "run_triangulation" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_triangulation" says "run": a word for a change.

    What this check looks for
  • The server marks "run_peer_benchmark" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_peer_benchmark" says "run": a word for a change.

    What this check looks for
  • The server marks "run_ai_visibility" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_ai_visibility" says "run": a word for a change.

    What this check looks for
Show 22 more findings
  • The server marks "run_ai_visibility_quick" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_ai_visibility_quick" says "run": a word for a change.

    What this check looks for
  • The server marks "run_disclosure_check" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_disclosure_check" says "run": a word for a change.

    What this check looks for
  • The server marks "run_regulatory_clock" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_regulatory_clock" says "run": a word for a change.

    What this check looks for
  • The server marks "run_deal_activity" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_deal_activity" says "run": a word for a change.

    What this check looks for
  • The server marks "run_demand_check" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_demand_check" says "run": a word for a change.

    What this check looks for
  • The server marks "run_answer_share" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_answer_share" says "run": a word for a change.

    What this check looks for
  • The server marks "run_answer_share_quick" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_answer_share_quick" says "run": a word for a change.

    What this check looks for
  • The server marks "run_share_of_search" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_share_of_search" says "run": a word for a change.

    What this check looks for
  • The server marks "run_premium_check" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_premium_check" says "run": a word for a change.

    What this check looks for
  • The server marks "run_angle_demand" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_angle_demand" says "run": a word for a change.

    What this check looks for
  • The server marks "run_channel_mix" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_channel_mix" says "run": a word for a change.

    What this check looks for
  • The server marks "run_claim_check" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_claim_check" says "run": a word for a change.

    What this check looks for
  • The server marks "run_fad_check" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_fad_check" says "run": a word for a change.

    What this check looks for
  • The server marks "run_revenue_mix" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_revenue_mix" says "run": a word for a change.

    What this check looks for
  • The server marks "run_launch_lift" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_launch_lift" says "run": a word for a change.

    What this check looks for
  • The server marks "run_comment_themes" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_comment_themes" says "run": a word for a change.

    What this check looks for
  • The server marks "run_customer_concentration" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_customer_concentration" says "run": a word for a change.

    What this check looks for
  • The server marks "run_write_downs" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_write_downs" says "run": a word for a change.

    What this check looks for
  • The server marks "run_cost_squeeze" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_cost_squeeze" says "run": a word for a change.

    What this check looks for
  • The server marks "run_trade_concentration" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_trade_concentration" says "run": a word for a change.

    What this check looks for
  • The server marks "run_series_forecast" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_series_forecast" says "run": a word for a change.

    What this check looks for
  • The server marks "run_threshold_odds" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_threshold_odds" says "run": a word for a change.

    What this check looks for

History

Changes to its tool definitions and to its remote endpoint, newest first.

  1. PostureRemote posture changed. Tools that change things, listed without credentials: 18 -> 29; RFC 9728 protected resource metadata: none -> https://cultmath.com/.well-known/oauth-protected-resource/mcp; Authorization servers: none -> https://cult…View diff
    • Tools that change things, listed without credentials: 18 to 29
    • RFC 9728 protected resource metadata: none to https://cultmath.com/.well-known/oauth-protected-resource/mcp
    • Authorization servers: none to https://cultmath.com
    • Dynamic client registration: not observed to yes
    • Client ID metadata documents: not observed to no
  2. Tools22 added, 6 changed.+22 added~6 changedView diff

Install

Pinned to version 1.3.0. Values in angle brackets are yours to fill in.

claude mcp add --transport http decision-briefs https://cultmath.com/mcp
claude mcp add --transport http decision-briefs https://cultmath.com/mcp

Run in a terminal.

Claude Code permissions

Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.

.claude/settings.json
{
  "permissions": {
    "allow": [
      "mcp__decision-briefs__list_decision_briefs",
      "mcp__decision-briefs__get_decision_brief",
      "mcp__decision-briefs__search_cultmath",
      "mcp__decision-briefs__list_decisions",
      "mcp__decision-briefs__get_decision_evidence"
    ],
    "ask": [
      "mcp__decision-briefs__buy_decision_evidence",
      "mcp__decision-briefs__get_document",
      "mcp__decision-briefs__run_filing_trend",
      "mcp__decision-briefs__run_regulatory_trend",
      "mcp__decision-briefs__run_spend_trend",
      "mcp__decision-briefs__run_triangulation",
      "mcp__decision-briefs__run_peer_benchmark",
      "mcp__decision-briefs__run_ai_visibility",
      "mcp__decision-briefs__run_ai_visibility_quick",
      "mcp__decision-briefs__run_disclosure_check",
      "mcp__decision-briefs__run_regulatory_clock",
      "mcp__decision-briefs__run_deal_activity",
      "mcp__decision-briefs__run_demand_check",
      "mcp__decision-briefs__run_answer_share",
      "mcp__decision-briefs__run_answer_share_quick",
      "mcp__decision-briefs__run_share_of_search",
      "mcp__decision-briefs__run_premium_check",
      "mcp__decision-briefs__run_angle_demand",
      "mcp__decision-briefs__run_channel_mix",
      "mcp__decision-briefs__run_claim_check",
      "mcp__decision-briefs__run_fad_check",
      "mcp__decision-briefs__run_revenue_mix",
      "mcp__decision-briefs__run_launch_lift",
      "mcp__decision-briefs__run_comment_themes",
      "mcp__decision-briefs__run_customer_concentration",
      "mcp__decision-briefs__run_write_downs",
      "mcp__decision-briefs__run_cost_squeeze",
      "mcp__decision-briefs__run_trade_concentration",
      "mcp__decision-briefs__run_series_forecast",
      "mcp__decision-briefs__run_threshold_odds"
    ],
    "deny": []
  }
}

Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as decision-briefs. Review it before you use it.

Capability tags behind 31 rules
  • mcp__decision-briefs__buy_decision_evidence untrusted-input
  • mcp__decision-briefs__get_decision_evidence untrusted-input
  • mcp__decision-briefs__get_document private-data, untrusted-input, outbound
  • mcp__decision-briefs__run_filing_trend untrusted-input
  • mcp__decision-briefs__run_regulatory_trend untrusted-input
  • mcp__decision-briefs__run_spend_trend untrusted-input
  • mcp__decision-briefs__run_triangulation untrusted-input
  • mcp__decision-briefs__run_peer_benchmark untrusted-input
  • mcp__decision-briefs__run_ai_visibility untrusted-input
  • mcp__decision-briefs__run_ai_visibility_quick untrusted-input
  • mcp__decision-briefs__run_disclosure_check untrusted-input
  • mcp__decision-briefs__run_regulatory_clock untrusted-input
  • mcp__decision-briefs__run_deal_activity untrusted-input
  • mcp__decision-briefs__run_demand_check untrusted-input
  • mcp__decision-briefs__run_answer_share untrusted-input
  • mcp__decision-briefs__run_answer_share_quick untrusted-input
  • mcp__decision-briefs__run_share_of_search untrusted-input
  • mcp__decision-briefs__run_premium_check untrusted-input
  • mcp__decision-briefs__run_angle_demand untrusted-input
  • mcp__decision-briefs__run_channel_mix untrusted-input
  • mcp__decision-briefs__run_claim_check untrusted-input
  • mcp__decision-briefs__run_fad_check untrusted-input
  • mcp__decision-briefs__run_revenue_mix untrusted-input
  • mcp__decision-briefs__run_launch_lift untrusted-input
  • mcp__decision-briefs__run_comment_themes untrusted-input, outbound
  • mcp__decision-briefs__run_customer_concentration untrusted-input
  • mcp__decision-briefs__run_write_downs untrusted-input
  • mcp__decision-briefs__run_cost_squeeze untrusted-input
  • mcp__decision-briefs__run_trade_concentration untrusted-input
  • mcp__decision-briefs__run_series_forecast untrusted-input
  • mcp__decision-briefs__run_threshold_odds untrusted-input

Pin

Keep the decision: fail CI when the tools change, follow every change, or show a badge.

Pin in CI

Fail a build when the tools you reviewed are not the tools the server serves today.

curl -s "https://radar.anlyon.com/api/v1/servers/com.cultmath/decision-briefs/verify?hash=42c4ae4979bde8a585e2382bb4ee47f684dc066f34459a7606333d37795d7f0b" | jq -e '.changed == false'

Pinned hash 42c4ae4979bd, from Nov 14, 2026, 16:17 UTC.

Watch

Get told when its tools change: a new tool, a new description, a new class.

https://radar.anlyon.com/servers/com.cultmath/decision-briefs/changes.xml

An Atom feed. Works in any feed reader or Slack's RSS app.

Badge

Anlyon Radar badge for com.cultmath/decision-briefs
[![Anlyon Radar](https://radar.anlyon.com/badge/com.cultmath/decision-briefs.svg)](https://radar.anlyon.com/servers/com.cultmath/decision-briefs)

Source: MCP Registry. Synced Nov 14, 2026, 16:17 UTC.

CULT+MATH Decision Briefs MCP server: questions and answers

Answered from the scan above, not from the publisher's description.

Is the CULT+MATH Decision Briefs MCP server safe?

No scan can promise that. What its definitions show: 35 tools, 29 of which can change, send or delete something, and 28 findings from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.

Can the CULT+MATH Decision Briefs MCP server delete data?

No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.

Can the CULT+MATH Decision Briefs MCP server do things that cannot be undone?

No tool is classed as irreversible: none of them sends, pays, publishes or merges, going by their definitions.

What tools does the CULT+MATH Decision Briefs MCP server have?

35 tools: 29 write, 6 read. buy_decision_evidence, run_filing_trend, run_regulatory_trend, run_spend_trend, run_triangulation, run_peer_benchmark, run_ai_visibility, run_ai_visibility_quick and 27 more.

Does the CULT+MATH Decision Briefs MCP server need sign-in or OAuth?

Its remote endpoint listed its tools to an anonymous client, without sign-in.

Does the CULT+MATH Decision Briefs MCP server send data to outside services?

2 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): get_document and run_comment_themes.

Has the CULT+MATH Decision Briefs MCP server changed its tools?

Yes, once since Anlyon Radar started tracking it, most recently on Nov 14, 2026. Every change is on this page and in its Atom feed.