Skip to main content
Glama
Sign In

Forecast Calibration Ledger: agent self-forecasts, resolved and scored

by com.textforagents
OverviewDiscussionsAdmin

Server Details

Signed agent self-forecasts, resolved and scored per cohort, with a self-killing trust weight.

Ownership verified
Status
Healthy
Uptime
100.0% over the last 24 hours
Last Tested
Transport
Streamable HTTP · MCP 2025-03-26
URL
ChatGPT
–
Claude
–

TDQS

B3.2/5.0

Scored across 4 tools

Disambiguation5/5

The four tools map to distinct stages of the forecast lifecycle: recording, resolving, querying calibration metrics, and checking trust/kill state. Although calibration and weight both query forecaster+cohort status, their descriptions clearly separate aggregate scoring from trust-state lookup.

Naming Consistency4/5

All tools share the consistent fcl_ prefix, making the set easy to recognize. The suffixes mix nouns and verbs (calibration/weight vs. forecast/resolve), which is a minor inconsistency but still readable and predictable.

Tool Count5/5

Four tools is well-scoped for a focused forecast-calibration ledger. Each tool earns its place by covering a core lifecycle operation without unnecessary surface area.

Completeness4/5

The surface covers the essential lifecycle: record a forecast, resolve it, query calibration, and inspect trust weight. Minor gaps remain, such as listing or retrieving individual forecasts and explicit batch-resolution support, but core workflows are covered.

Available Tools

4 tools
fcl_calibrationBInspect

Calibration query: Brier, skill vs climatology +/- SEM, bins, cost/duration error, trust weight ($0.005/query; included with fleet plan). Requires api_token, forecaster, cohort; optional x_payment.

Inputs: cohort (required), api_token (required), forecaster (required), x_payment.

fcl_forecastBInspect

Record a signed forecast bound to run_id. Requires api_token (from POST /register), forecaster, cohort, run_id, p, cost_est, duration_est.

Inputs: p (required), cohort (required), run_id (required), cost_est (required), +3 more.

fcl_resolveCInspect

Resolve a forecast against its run record ($0.01/forecast; one free 100-forecast batch per operator, ever; fleet plan includes 10,000). Requires api_token, run_id, success, cost_actual, duration_actual, run_record_source; optional x_payment.

Inputs: run_id (required), success (required), api_token (required), cost_actual (required), +3 more.

fcl_weightCInspect

Trust weight / kill state for your forecaster+cohort (free). Requires api_token, forecaster, cohort.

Inputs: cohort (required), api_token (required), forecaster (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • numguardprivate

      ipezygj
      A
      license
      B
      quality
      B
      maintenance
      Verify a number before an agent asserts it — a Deflated Sharpe Ratio for backtest, plus eval-gap, subset-win, and judge-bias checks, with signed receipts anyone can verify offline.
      34
      MIT

    • p1b

      rui08984-dot
      A
      license
      Not graded
      quality
      B
      maintenance
      Enables users to record probabilistic judgments in an immutable ledger, automatically score them against outcomes, and analyze systematic biases across six classification layers. It supports natural-language interaction for logging predictions, checking randomness, and auditing data integrity while avoiding advice or replacing user judgment.
      Apache 2.0

    • agiscorecard

      f-tiger
      F
      license
      Not graded
      quality
      B
      maintenance
      Auditable AGI-2027 evidence: eight graded Situational Awareness predictions with pre-registered flip conditions, a 0-100 Thesis Tracker with full score history, and a public market-call ledger where misses stay published. Free, no auth, CC BY 4.0.
      -
    Try in Browser
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    Glama MCP Gateway

    Add one secure layer between your agents and this server.

    Resources