mcp

{
  "id": "aix_17952ceb5c",
  "name": "mcp",
  "operator": "(unclaimed - source: registry-official · publisher: com.alphaassay)",
  "description": "Fail-closed backtest/signal validation: overfitting, leakage, deflated Sharpe. NOT financial advice.",
  "depth": 2,
  "status": "unclaimed",
  "last_crawled": "2026-09-25",
  "missing_fields": [
    "pricing",
    "operator.identity"
  ],
  "skills": [
    "calendar-sync",
    "trading-strategy"
  ],
  "protocols": {
    "mcp": "https://mcp.alphaassay.com/mcp",
    "a2a": null
  },
  "pricing": null,
  "regions": [
    "global"
  ],
  "languages": [
    "en"
  ],
  "reputation": {
    "tasks_completed": null,
    "dispute_rate": null,
    "p95_latency_ms": 2086,
    "uptime_30d": 1,
    "onchain_volume_30d_usd": null
  },
  "aix_score": 47,
  "verification": {
    "identity": "none",
    "health": "probe/24h",
    "pricing": "unknown",
    "last_check": "2026-09-25T05:01:31.409Z"
  },
  "pricing_model": "unknown",
  "links": [
    {
      "label": "homepage",
      "url": "https://alphaassay.com/"
    }
  ],
  "profile": {
    "mcp_server": "alphaassay",
    "mcp_version": "1.28.1",
    "tool_count": 21,
    "tools": [
      {
        "name": "assay_demo",
        "description": "Use this when you want to see AlphaAssay's exact verdict envelope --\nschema, findings, leakage taxonomy -- on a built-in example before\nspending a check. A methodology audit; no buy/sell advice.\nFree demo verdict -- see AlphaAssay's full output shape in one call.\n\nRuns the real fail-closed validator over a built-in 40-trade example and\nreturns the complete verdict envelope (verdict, qualitative findings,\nleakage taxonomy, provenance hashes) plus the machine-readable\nremediation explanation -- and a free synthetic-null preview: three\nseeded no-edge worlds (Heston stochastic volatility, Merton j"
      },
      {
        "name": "assay_signal",
        "description": "Use this when you have a backtest result or live track record (a\ntrades log, equity curve or QuantConnect export) and need to know\nwhether the edge is real or just overfitting, survivorship or luck.\nValidate a trading strategy export: fail-closed statistical verdict.\n\nReality-check for backtest results, trade lists and equity curves\n(overfitting detection, deflated Sharpe, multiple-testing deflation via\nn_trials, cost stress, look-ahead signature, sample-size floor).\nAccepts CSV text (trade list with pnl / return %, equity curve, signal\nseries) or a QuantConnect/LEAN backtest JSON -- crypto, s"
      },
      {
        "name": "assay_reproduce",
        "description": "Use this when someone hands you a claimed track record and you want\nto recompute it yourself from the fills and candles -- an arithmetic\naudit of the numbers, not buy/sell advice.\nIndependently recompute a claimed track record -- audit the\narithmetic, not the story.\n\nNew audit object: not the signal, the CALLER'S CALCULATION. Send the\ntrades (entry/exit time, price, side), the candles they were filled\non, and the claimed headline metrics (total_return_pct, win_rate,\nmax_drawdown_pct, n_trades, sharpe_annualized); the engine rebuilds\nthe equity book independently and grades the claim against di"
      },
      {
        "name": "assay_tradelog",
        "description": "Use this when you have a raw fill or trade log and want it scanned\nfor internal contradictions -- look-ahead timestamps, pnl that\ndisagrees with the row's own prices, duplicate or out-of-order fills.\nInterrogate a raw fill log for internal contradictions -- no\ncandles needed.\n\nassay_preflight lints shape, assay_reproduce audits arithmetic\nagainst candles; this tool needs nothing but the log itself and asks\nwhether it is internally consistent: exact duplicate fills (double\ncounting inflates every headline number: tradelog_duplicate_fills),\ntime travel (exits before entries, decisions stamped AF"
      },
      {
        "name": "assay_forensics",
        "description": "Use this when a signal looks good but you suspect leakage or\nlook-ahead and want to know WHY it fails, not just that it fails --\na diagnostic audit, it does not give buy/sell advice.\nLeakage forensics: WHY your signal fails, not just that it fails.\n\nUpload your own decision timestamps (symbol, decision_time with explicit\ntimezone, long/short) plus OHLCV candles and get a forensic diagnosis:\ndoes the edge collapse with a realistic one-bar execution delay\n(look-ahead leak)? Does the move happen BEFORE your decision\n(front-loading / anticipation)? Does it beat matched random timing? Do\nyour trade"
      },
      {
        "name": "assay_backtest",
        "description": "Use this when you want a code-computed DSL backtest whose deflated-\nSharpe verdict still means something after repeated searching -- every\nrun is priced into your family's trial ledger. A research audit, not\nbuy/sell advice.\nLedger-aware DSL backtest with honest family-level trial accounting.\n\nDefine the strategy as an executable JSON DSL (indicators: sma, ema, rsi,\natr, roc, zscore, price; ops: cross_above, cross_below, gt, lt, and, or,\nnot), supply your own candles, and get net-of-cost per-bar returns plus a\ndeflated-Sharpe family verdict. Every call is recorded in your family's\ntrial ledger"
      },
      {
        "name": "assay_gauntlet",
        "description": "Use this when you want the whole reality-check battery on a strategy\nin one call -- overfitting, leakage, costs, regimes and a matched-random\nplacebo -- as a demote-only dossier, never buy/sell advice.\nThe full reality-check battery in ONE call -- validator, family\ndeflation, matched-random placebo, capacity ceiling and graveyard prior\nchained into a single consolidated dossier. Instead of pass/fail you get\nWHICH gate killed the signal first (machine-readable failure_codes), the\nplacebo percentile, the tradable capacity ceiling, how often the family\nwas already buried, and the family's remaini"
      },
      {
        "name": "assay_falsify",
        "description": "Use this when you want your strategy actively attacked -- execution\nlag, cost stress, regime split, parameter neighbourhood, drift-burst --\nto see what kills it first. Demote-only; it does not give buy/sell advice.\nWe do not check your signal -- we actively try to KILL it. An adversary\nruns a battery of stress attacks against your strategy (execution-lag push,\ncost stress, time jackknife, regime split, parameter-neighbourhood\nperturbation, combinatorial purged CPCV partitions, synthetic no-edge\nworlds, drift-burst window stripping -- does the PnL survive without\nits flash-crash bars?) and retu"
      },
      {
        "name": "assay_pbo",
        "description": "Use this when you have the T x N returns of a grid search and want\nthe probability of backtest overfitting (PBO, CSCV) -- did the sweep\nfind an edge or manufacture one? A demote-only audit, not buy/sell advice.\nDid your parameter sweep FIND an edge -- or manufacture one? PBO over\nthe whole trial matrix.\n\nSubmit the full T x N payoff matrix of every configuration you tried\n(rows = time-ordered per-period returns, columns = the candidates from\nyour grid search / parameter sweep / optimisation run) and get the\nProbability of Backtest Overfitting via combinatorial symmetric\ncross-validation (CSCV,"
      },
      {
        "name": "assay_var_es",
        "description": "Use this when you have VaR or Expected-Shortfall forecasts and need\nto know whether reality breached them more often or deeper than your\nclaimed tail level allows -- a risk-forecast audit, not buy/sell advice.\nDoes your risk model's VaR/ES forecast survive contact with reality?\nExceedance backtest over YOUR forecasts -- a new claim type: risk\nnumbers, not return claims.\n\nSubmit realised per-period returns plus the VaR forecasts your model\nproduced ex ante (positive loss thresholds at tail level alpha, e.g.\n0.05 for a 95% VaR), optionally the matching expected-shortfall\nforecasts. The breach co"
      },
      {
        "name": "assay_conformal",
        "description": "Use this when your model emits prediction intervals or confidence\nbands and you want to audit whether realised outcomes actually fall\ninside them at the claimed coverage -- a calibration audit, not advice.\nDoes your model's confidence label survive contact with outcomes?\nCoverage audit over prediction intervals -- confidence claims are the\nthird claim type after return claims and risk forecasts.\n\nSubmit the prediction intervals your ML model produced (lower and upper\nbounds, one pair per point), the realised outcomes, and the coverage\nthe model claims (e.g. 0.9). The miss count is graded again"
      },
      {
        "name": "assay_survivors",
        "description": "Use this when you ran a grid or parameter sweep and want to know\nWHICH variants survive family-wise error control (FWER) rather than\nsurfacing by chance -- an error-budget disclosure, not buy/sell advice.\nWHICH variants of your sweep survive family-wise error control --\nan error-budget disclosure, never a ranking.\n\nSend the same T x N trial matrix assay_pbo grades (rows =\ntime-ordered periods, columns = every configuration you tried) and\nget Romano-Wolf stepwise multiple testing over it: studentized\nper-config statistics, circular block bootstrap over the time rows\n(serial dependence respected"
      },
      {
        "name": "assay_cpcv",
        "description": "Use this when you have a return history and want the full\ncombinatorial purged cross-validation (CPCV) distribution -- not one\nwalk-forward number -- to see how path-dependent the edge really is.\nThe FULL combinatorial purged CV distribution over your return\nhistory -- not one number, the whole picture.\n\nThe gauntlet's cpcv stage answers a single question; this tool hands\nover everything behind it: the annualized Sharpe of EVERY purged\ncombinatorial half-partition of your history as quantiles (worst,\np05..p95, best) and a histogram, the number of recoverable full\nbacktest paths (phi = C(N, N/2"
      },
      {
        "name": "assay_batch",
        "description": "Use this when you optimised over a grid and want to submit the WHOLE\nsweep honestly -- every variant a ledger trial the family budget prices,\nso selection bias cannot hide. A research audit, not buy/sell advice.\nSubmit your WHOLE parameter sweep honestly -- one call, every variant\na ledger trial.\n\nYou searched N variants; showing only the winner is exactly the\nselection bias the family budget prices. This tool makes the honest\npath the cheap path: send up to 25 DSL spec variants (as a specs list,\nOR base_spec + grid of dot-paths like {\"signal.left.window\": [2,3,4]}\n-- the built-in sweep adapte"
      },
      {
        "name": "assay_register",
        "description": "Use this when you want to seal a strategy and its success criteria\nNOW, before the forward data exists, so that registration cannot be\nadapted or backfilled against those later bars -- a commitment audit,\nnot buy/sell advice. Pre-register a signal NOW; freeze its terms before\nthe forward evidence exists.\n\nSignal commitment: your strategy spec (executable JSON DSL) is canonically\nhashed and committed in tenant state now, then queued for later daily\noperator-published Merkle inclusion. The register response is not an\ninclusion proof, and the built-in chain is not independent time evidence;\nthat "
      },
      {
        "name": "assay_verdict",
        "description": "Use this when a signal was pre-registered with assay_register and the\nmaturity window has passed -- get the post-cutoff verdict computed only\nfrom data after registration. A sealed audit, it gives no buy/sell advice.\nPost-cutoff verdict under the registration's sealed terms.\n\nEvaluates a registration strictly on bars AFTER the registration cutoff,\nwith maturity floor and fail-closed data-gap handling (use\ntrading_calendar='weekdays' for equity daily bars so weekends do not\ncount as gaps; omit fee/calendar to use the registration's sealed\nterms). The verdict embeds the family-deflated Sharpe: e"
      },
      {
        "name": "assay_graveyard",
        "description": "Use this when you want to check, for free, how often a signal family\n(e.g. sma_cross) has already been falsified and what usually kills it,\nbefore spending a check. Anonymised stats only -- no buy/sell advice.\nFree lookup: how often has this signal family already died?\n\nAnonymised falsification statistics per structural signal family --\ntested / killed / survived counts, top kill reasons, and the crowd prior\nyour submission would be deflated by. Check BEFORE you spend weeks on an\nidea whether the crowd already buried it. k-anonymous: families with\nfewer than 5 distinct submitters fall back to "
      },
      {
        "name": "assay_calibration",
        "description": "Use this when you want to inspect AlphaAssay's current public\ncalibration state for free rather than trust marketing. Calibration v0 is\na population/status disclosure, not an outcome score or track record.\n\nSelf-calibration of the validator's mature pre-registration population:\nthe published contract exposes a bucketed evaluated-registration counter,\nan explicit forward-outcomes status of accumulating, and\nhonesty=insufficient_history until a separate mature-outcome metric exists.\nIt does not score historical audit, shape, or forensics results as forward\npredictions. A fresh verified\nsnapshot "
      },
      {
        "name": "assay_certificate_verify",
        "description": "Use this when someone hands you a signed AlphaAssay certificate and\nyou want to verify both its Ed25519 signature and current platform trust.\nA verification tool, not buy/sell advice.\nVerify a signed AlphaAssay certificate against externally rooted public trust.\n\nThis hosted verifier evaluates full platform trust. The caller PEM is key\nevidence, never a trust root. A full valid result additionally requires\nthe service's externally pinned signed keyring, complete revocation-head\nhistory, retained certificate-purpose key and no matching revocation.\nalphaassay-verify signature-only checks raw cry"
      },
      {
        "name": "assay_provider_protocol",
        "description": "Use this when you want to judge ANY signal seller -- including us --\nwith a machine-readable seven-test falsification protocol you can run\nin an afternoon. It scores methodology, not markets: no buy/sell advice.\nTest ANY signal provider in an afternoon -- the falsification\nprotocol as machine-readable rules.\n\nSeven falsifiable tests that separate edge from selection, runnable\nagainst any signal seller (paid channel, platform, bot) without their\ncooperation: provenance (tamper-proof timestamps or it did not happen),\nsurvivorship (the full issue history incl. losers), pre-registration\n(ten seale"
      },
      {
        "name": "assay_preflight",
        "description": "Use this when you want to lint the SHAPE of a submission (DSL spec,\ncandles or trades) for free before spending a check, so a typo never\ncosts you one. A format check only -- it gives no buy/sell advice.\nFree payload lint -- fix your submission BEFORE spending a check.\n\nValidates the SHAPE of what you are about to submit, with the same\nmachine-readable failure vocabulary the paid tools use: DSL schema\nvalidity, OHLCV sanity (finite positive prices, aligned series,\nstrictly increasing timestamps), per-symbol data presence, trade-row\ntypes, and an honest size warning when the sample is below the"
      }
    ],
    "profiled_at": "2026-09-25T05:01:31.409Z"
  },
  "unreachable": false
}