io.github.inity13由 GitHub 帳號 inity13 發布v1.0.1更新於 Oct 27, 2026
Deterministic reasoning stack for AI agents: simulate, decide & compute, plus cross-domain tools.
確定性的推理工具組，讓助理執行假設情境模擬、依加權準則排序選項，並精確計算財務與業務指標。
Web executable，透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。
把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式，或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南，先向你展示發布者、端點和登入方式，經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。
閱讀 https://sourceweft.com/mcp/io.github.inity13%2Fagentstack-mcp/install.md，並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前，先告訴我它的發布者、端點和登入方式，等我確認。不要把 API 金鑰寫進命令或我會提交的檔案；問我金鑰應從哪裡取得。把它新增到你客戶端的 mcpServers 設定中。
{
"mcpServers": {
"agentstack": {
"type": "http",
"url": "https://agentstack-mcp.pages.dev/mcp"
}
}
}One deterministic reasoning stack for AI agents — behind a single Model Context Protocol (MCP) endpoint and one API key:
simulate (ScenarioSim) → decide (DecisionMatrix) → compute (PrecisionCalc)
plus cross-domain composite tools that chain the three. Every number runs
through decimal.js at 40-digit precision
(never floats), so identical inputs always produce byte-identical output.
The server is stateless — no database, no sessions.
AgentStack bundles three standalone MCP servers so an agent installs one server, uses one key, and can run reasoning that spans all three domains. It imports the same engines directly (no HTTP proxying) — zero added latency, no cascading failure.
Free tier: no key, 20 calls/day per IP. Paid plans (Starter $19/mo · 8,000/day,
Pro $59/mo · 80,000/day) return an API key sent as X-API-Key. One key covers
all three products and the composites — about half the cost of subscribing to them
separately. Landing page: https://agentstack-mcp.pages.dev.
Note on two deliberately-namespaced collisions:
sim_sensitivityvaries scenario inputs;decide_sensitivityvaries criteria weights. They are different operations — the prefix disambiguates them for the agent.
Deterministic end-to-end; each chains engines that no single server bundles:
Composite responses report the pipeline they ran and carry each stage's
structured output, so an agent can drill into any step.
Append ?profile= to the endpoint to load only a subset:
Namespaced tools return the exact envelope of their underlying product. Composite
tools add composite, pipeline, per-stage blocks, methodology, notes, and a
plain-language explanation. Errors are always structured data:
Example — plan_to_valuation (trimmed):
Why vendor the engines? So this repo is self-contained and reproducible. The
engines are the same deterministic modules that power the three standalone
servers; npm run sync-engines re-copies them from the sibling repos when an
upstream change lands.
Quick manual call:
Run the server locally over stdio with a single command — nothing to deploy:
Claude Desktop / any stdio MCP client (claude_desktop_config.json):
This is the same deterministic engine as the hosted server, running on your machine.
~/.cursor/mcp.jsonclaude_desktop_config.json (via mcp-remote).vscode/mcp.json~/.codeium/windsurf/mcp_config.jsonAdd "headers": { "X-API-Key": "ssak_live_..." } once you have a key.
Rank growth strategies on their projected outcomes:
Value a plan's MRR stream:
Stress-test the decision:
Runs fully free / private with no bindings or secrets (fails open on the free
tier). For billing: bind a KV namespace AGENTSTACK_KV, set PRICE_STARTER /
PRICE_PRO in [vars], and add Stripe secrets:
Routes: /mcp (Streamable HTTP), /checkout, /success, /portal, /webhook, /metrics.
Enforced in worker-src/billing.mjs: identify() reads X-API-Key /
Authorization: Bearer and falls back to per-IP free tier; consumeQuota() is a
KV daily counter (the single gating point). One key / one quota spans the whole
stack. Marked NOTE (auth) / NOTE (rate limiting) seams show where to swap in
JWT/mTLS or a token-bucket. None of it touches the deterministic math.
AgentStack is additive, not a replacement. The three products keep running on their own for single-domain use:
MIT — see LICENSE.
來源：README.md，提交 7f659dd
Discovery: the three namespaces (sim_*, decide_*, calc_*), the cross-domain composite tools, the available ?profile= filters, and links to the standalone servers. Call this first to see everything AgentStack exposes. No parameters.
無需輸入參數。
Aggregated health/status for the whole stack (all three engines + composites). No parameters.
無需輸入參數。
SIMULATE. Deterministic what-if projection from a template (saas_growth, pricing_change, churn_impact, cost_reduction, hiring_plan, cash_runway, unit_economics, marketing_funnel, compound_growth) or a free-form 'metrics' model. Returns per-period projections, key_results, assumptions_used, methodology, and an explanation.
inputshorizonmetricstemplateperiod_labelSIMULATE. Vary one or more scenario inputs and show the impact on a target output metric (one-at-a-time), with elasticity + most-influential ranking. Requires 'template' and 'variable' (or 'variables').
maxminstepsinputsvalueshorizontemplate必填variablevariablesvariationperiod_labeltarget_metricSIMULATE. Solve for the scenario input value required to make an output metric hit a target value (deterministic bisection). Requires 'template', 'solve_for', 'target_value'.
boundsinputshorizontemplate必填solve_for必填period_labeltarget_value必填target_metricSIMULATE. Run 2-3 scenarios and compare their key_results side by side with deltas vs the first (baseline). Optional 'compare_metric' + 'goal' (max|min) picks a winner.
goalhorizonscenarios必填compare_metricinclude_projectionsSIMULATE. List every scenario template (inputs, defaults, outputs) plus the custom-model format and period labels. No parameters.
無需輸入參數。
DECIDE. Rank named options against weighted criteria and return the winner, full ranking, per-criterion breakdowns, methodology, weights, and an explanation. Provide options, criteria [{name, weight, direction}], and a scores matrix. method: weighted_sum (default) | weighted_product | topsis.
methodscores必填options必填criteria必填DECIDE. Return the full normalized scored matrix (per-option, per-criterion) + ranking when scores are supplied separately, without the winner narrative.
methodscores必填options必填criteria必填DECIDE. Test how robust the decision winner is to changes in CRITERIA WEIGHTS (distinct from sim_sensitivity, which varies scenario inputs). Sweeps each weight +/-variation and reports a robustness score + flip points.
stepsmethodscores必填options必填criteria必填variationDECIDE. Head-to-head comparison of exactly two options with per-criterion win counts and margin. Pass option_a/option_b (or a 2-element options array), criteria, and scores.
methodscores必填optionscriteria必填option_aoption_bDECIDE. List the scoring methods (weighted_sum, weighted_product, topsis) with normalization details and when to use each. No parameters.
無需輸入參數。
COMPUTE. Exact business/SaaS/finance metric: ltv, cac, ltv_cac_ratio, payback_period_months, contribution_margin, gross_margin, churn_rate, mrr_growth_rate, arr, break_even_units, nrr, grr, rule_of_40, magic_number. Rates/margins are decimals (0.05=5%). Call calc_list_metrics for schemas.
metric必填params必填currencyCOMPUTE. List every supported metric with descriptions and required/optional params. No parameters.
無需輸入參數。
COMPUTE. Convert between major currencies (USD, EUR, GBP, JPY, CAD, AUD, CHF, CNY, INR) with Decimal precision. Static offline table by default; live/historical ECB rates via date/live=true.
dateliveamount必填to_currency必填from_currency必填COMPUTE. Business-day arithmetic honoring weekends + regional holidays. operation: add_business_days | count_business_days | next_business_day | previous_business_day. region: US | UK | EU | NONE.
daysregionend_dateoperation必填start_date必填custom_holidaysCOMPUTE. Compound-interest/growth math. operation: future_value | present_value | cagr. rate is annual decimal; compounding: daily|weekly|monthly|quarterly|semiannually|annually|continuous.
rateyearscurrencyend_valueoperation必填begin_valuecompoundingfuture_valuepresent_valueCOMPUTE. Net Present Value (discounted cash flow). NPV = sum(CF_t/(1+rate)^t); cashflows[0] is period 0 (usually the negative outlay).
rate必填currencycashflows必填COMPUTE. Internal Rate of Return: per-period rate where NPV=0 (Newton + bisection). Requires a sign change in cashflows.
guesscashflows必填COMPUTE. Level-payment loan: monthly payment, total interest, payoff, and (optional) full schedule.
currencyprincipal必填annual_rate必填term_months必填extra_paymentinclude_scheduleCOMPUTE. Asset depreciation schedule. method: straight_line | declining_balance | sum_of_years_digits.
cost必填method必填currencysalvage_value必填useful_life_years必填COMPOSITE (simulate -> compute). Project a scenario, take a per-period cash-flow line from its projections ('cashflow_metric', e.g. 'mrr' or 'net_burn'), and value it exactly: NPV at a discount 'rate', IRR, and undiscounted total. Optional 'initial_investment' becomes the period-0 outflow (needed for IRR). Combines ScenarioSim + PrecisionCalc.
rate必填inputshorizonmetricscurrencytemplate必填period_labelcashflow_metric必填initial_investmentCOMPOSITE (simulate -> decide). Project each option as its own scenario, then rank the options against weighted criteria drawn from the scenario OUTCOMES. Provide a base 'template', an 'options' array ([{name, inputs}]), and 'criteria' ([{metric, weight, direction}]) where each metric is a scenario key_result. Combines ScenarioSim + DecisionMatrix.
inputsmethodhorizonoptions必填criteria必填templateperiod_labelCOMPOSITE (simulate x decide). Take an options-vs-scenarios decision and stress ONE scenario assumption across a range applied to every option; report how often the baseline winner survives (robustness) and where it flips. Same args as evaluate_options_with_scenarios plus 'stress': {variable, variation, steps}.
inputsmethodstress必填horizonoptions必填criteria必填templateperiod_label已滿足適用檢查項的 100%
該評分衡量此收錄條目的完整程度，依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價，也不表示伺服器安全或目前可用：可連通性是帶日期的檢查結果，評分也不包含安裝量或使用者評價。
https://agentstack-mcp.pages.dev/mcp{ "mcpServers": { "agentstack": { "type": "http", "url": "https://agentstack-mcp.pages.dev/mcp" } } }{ "mcpServers": { "agentstack": { "url": "https://agentstack-mcp.pages.dev/mcp?profile=finance" } } }{ "status": "error", "error": { "type": "unknown_metric", "message": "...", "hint": "..." } }{ "status": "success", "composite": "plan_to_valuation", "pipeline": ["scenariosim.run_scenario", "precisioncalc.net_present_value", "precisioncalc.internal_rate_of_return"], "scenario": { "template": "saas_growth", "horizon": 12, "key_results": { "ending_mrr": 26982.1 } }, "cashflow_metric": "mrr", "cashflows": [-150000, 14040, 15600, "..."], "valuation": { "discount_rate": 0.01, "npv": 11157.386023, "irr": 0.02504, "value_creating": true }, "explanation": "Projecting 'saas_growth' over 12 month(s) and valuing the 'mrr' stream at 1%/period gives an NPV of ..."}agentstack-mcp/├── worker-src/│ ├── index.mjs # Cloudflare Pages Function: MCP transport, namespacing, profiles, billing│ ├── composites.mjs # the 3 cross-domain tools (simulate -> decide -> compute)│ ├── billing.mjs # Stripe + KV: one key / one quota for the whole stack│ └── engines/ # vendored, unmodified product engines (100% deterministic)│ ├── scenariosim.mjs│ ├── decisionmatrix.mjs│ └── precisioncalc.mjs├── server.mjs # local stdio MCP server (same engines + composites; AGENTSTACK_PROFILE env)├── site/ # landing page + mcp.json + llms.txt + _worker.js (built)├── tests/composites.test.mjs├── examples/agent_example.mjs├── scripts/sync-engines.sh # re-vendor engines from the sibling repos├── package.json · wrangler.toml · server.json · smithery.yaml · Dockerfile · LICENSE└── README.mdgit clone <your-fork> agentstack-mcp && cd agentstack-mcpnpm installnpm test # composite-logic tests (no network)npm run dev # -> http://127.0.0.1:8788/mcpnode examples/agent_example.mjs # hosted demonode examples/agent_example.mjs http://127.0.0.1:8788node server.mjs # dependency-light stdio server (AGENTSTACK_PROFILE=finance to subset)curl -s http://127.0.0.1:8788/mcp \ -H 'content-type: application/json' -H 'accept: application/json, text/event-stream' \ -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"list_capabilities","arguments":{}}}'npx -y agentstack-mcp{ "mcpServers": { "agentstack": { "command": "npx", "args": ["-y", "agentstack-mcp"] } } }{ "mcpServers": { "agentstack": { "url": "https://agentstack-mcp.pages.dev/mcp" } } }{ "mcpServers": { "agentstack": { "command": "npx", "args": ["-y", "mcp-remote", "https://agentstack-mcp.pages.dev/mcp"] } } }{ "servers": { "agentstack": { "type": "http", "url": "https://agentstack-mcp.pages.dev/mcp" } } }{ "mcpServers": { "agentstack": { "serverUrl": "https://agentstack-mcp.pages.dev/mcp" } } }{ "name": "evaluate_options_with_scenarios", "arguments": { "template": "saas_growth", "horizon": 12, "options": [ { "name": "Aggressive", "inputs": { "new_customers_per_period": 60, "churn_rate": 0.05 } }, { "name": "Lean", "inputs": { "new_customers_per_period": 20, "churn_rate": 0.02 } } ], "criteria": [ { "metric": "ending_mrr", "weight": 3, "direction": "benefit" }, { "metric": "total_churned_customers", "weight": 1, "direction": "cost" } ]} }{ "name": "plan_to_valuation", "arguments": { "template": "saas_growth", "horizon": 12, "inputs": { "new_customers_per_period": 60, "churn_rate": 0.05, "arpu": 60 }, "cashflow_metric": "mrr", "rate": 0.01, "initial_investment": 150000} }{ "name": "stress_test_decision", "arguments": { "template": "saas_growth", "horizon": 12, "options": [ { "name": "Aggressive", "inputs": { "churn_rate": 0.05 } }, { "name": "Lean", "inputs": { "churn_rate": 0.02 } } ], "criteria": [ { "metric": "ending_mrr", "weight": 3 } ], "stress": { "variable": "churn_rate", "variation": 0.5, "steps": 5 }} }npm installnpx wrangler loginnpm run deploy # esbuild -> site/_worker.js, then wrangler pages deploywrangler pages secret put STRIPE_SECRET_KEY --project-name agentstack-mcpwrangler pages secret put STRIPE_WEBHOOK_SECRET --project-name agentstack-mcp