NoiseFloor — market attention and evidence checks

by io.github.beepboop2025

Describe a metric change

did_it_change

Read-onlyIdempotent

Scan one chronological numeric metric for departures from its own preceding observations, using a descriptive rank-based Shiryaev-Roberts monitor. Use this for a change-review queue; choose real_or_sampling when a measurement denominator may explain the movement, market_assessment when source/freshness gates matter, or spectral_assessment for a multi-series correlation panel. The first eight readings warm up the reference; comparison starts at reading nine. An alarm resets the reference and starts warmup again. Returns state (warming_up, steady, watch, changed), monitoring_statistic, direction, zero-based change_points and effect (delta in input units, robust_z, percentile from 0 to 100). The evidence field is a compatibility alias, not a calibrated e-value or p-value; never use it to assert valid_evalues in which_metrics_matter. No alarm does not establish stability. Example: {"values":[10,11,9,10,12,10,9,11,20]}.

Input Schema

Name	Required	Description
`values`	Yes	One to 2048 finite readings in chronological order, oldest first, with consistent units and cadence. The last entry is the current reading; no dates or missing values are inferred.
`half_life`	No	Optional positive half-life measured in prior observations, not seconds. A reading this many positions older receives half the rank weight. Omit for equal weights; weighting remains descriptive, not calibrated.
`two_sided`	No	Default true monitors upward and downward departures, with watch/changed thresholds 200/1000. False monitors upward only, with thresholds 100/500; thresholds are descriptive.

Output Schema

Name	Required	Description
`n`	No	Number of supplied observations.
`error`	No	Validation or computation failure message; inspect MCP isError or the HTTP status.
`state`	No	Current descriptive monitor state.
`effect`	No	Latest reading compared with the current post-reset reference.
`method`	No	Monitor construction and one-sided/two-sided mode.
`reading`	No	Interpretation, including remaining warmup requirements.
`evidence`	No	Compatibility alias of monitoring_statistic; not an e-value.
`validity`	No	Descriptive assessment, not a probability or statistical certificate.
`watch_at`	No	Descriptive watch threshold.
`weighted`	No	Whether half-life weighting was requested.
`direction`	No	Direction driving the positive monitor statistic. Null means unavailable or not applicable, never zero.
`guarantee`	No	No statistical or market-performance guarantee is made.
`changed_at`	No	Descriptive alarm threshold.
`is_e_value`	No	The statistic is not a calibrated e-value.
`assumptions`	No	Ordered statements or identifiers.
`change_points`	No	Historical alarm indices.
`interpretation`	No	Limits on interpreting a threshold crossing.
`statistic_type`	No	Monitor family.
`monitoring_statistic`	No	Descriptive Shiryaev-Roberts statistic.
`effect_reference_size`	No	Preceding observations in the current reference.
`readings_since_change`	No	Observations after the most recent alarm. Null means unavailable or not applicable, never zero.

Schema Changelog

Changes observed during successful MCP inspections.

Changed8 schema fields changedOct 18, 2026

addedInput schema / properties / half_life / description

Added value: +"Optional positive half-life measured in prior observations, not seconds. A reading this many positions older receives half the rank weight. Omit for equal weights; weighting remains descriptive, not calibrated."

addedInput schema / properties / two_sided / default
```
Added value: +true
```

addedInput schema / properties / two_sided / description

Added value: +"Default true monitors upward and downward departures, with watch/changed thresholds 200/1000. False monitors upward only, with thresholds 100/500; thresholds are descriptive."

addedInput schema / properties / values / description

Added value: +"One to 2048 finite readings in chronological order, oldest first, with consistent units and cadence. The last entry is the current reading; no dates or missing values are inferred."

removedOutput schema / additionalProperties
```
Removed value: -true
```

addedOutput schema / anyOf

Added value: +[
+  {
+    "additionalProperties": false,
+    "properties": {
+      "assumptions": {},
+      "change_points": {},
+      "changed_at": {},
+      "direction": {},
+      "effect": {},
+      "effect_reference_size": {},
+      "evidence": {},
+      "guarantee": {},
+      "interpretation": {},
+      "is_e_value": {},
+      "method": {},
+      "monitoring_statistic": {},
+      "n": {},
+      "reading": {},
+      "readings_since_change": {},
+      "state": {},
+      "statistic_type": {},
+      "validity": {},
+      "watch_at": {},
+      "weighted": {}
+    },
+    "required": [
+      "n",
+      "state",
+      "evidence",
+      "monitoring_statistic",
+      "statistic_type",
+      "is_e_value",
+      "validity",
+      "weighted",
+      "assumptions",
+      "direction",
+      "effect",
+      "effect_reference_size",
+      "change_points",
+      "watch_at",
+      "changed_at",
+      "readings_since_change",
+      "reading",
+      "guarantee",
+      "interpretation",
+      "method"
+    ],
+    "type": "object"
+  },
+  {
+    "additionalProperties": false,
+    "description": "Existing tool-error payload; no successful result fields are implied.",
+    "properties": {
+      "error": {
+        "description": "Validation or computation failure message; inspect MCP isError or the HTTP status.",
+        "type": "string"
+      }
+    },
+    "required": [
+      "error"
+    ],
+    "type": "object"
+  }
+]

addedOutput schema / description

Added value: +"MCP change result; full per-reading states and evidence_series are deliberately omitted by this tool. Tool errors retain the separate error-message shape."

addedOutput schema / properties

Added value: +{
+  "assumptions": {
+    "description": "Ordered statements or identifiers.",
+    "items": {
+      "description": "One retained statement or identifier.",
+      "type": "string"
+    },
+    "type": "array"
+  },
+  "change_points": {
+    "description": "Historical alarm indices.",
+    "items": {
+      "description": "Zero-based observation index.",
+      "minimum": 0,
+      "type": "integer"
+    },
+    "type": "array"
+  },
+  "changed_at": {
+    "description": "Descriptive alarm threshold.",
+    "type": "number"
+  },
+  "direction": {
+    "anyOf": [
+      {
+        "description": "Direction driving the positive monitor statistic.",
+        "enum": [
+          "up",
+          "down"
+        ],
+        "type": "string"
+      },
+      {
+        "type": "null"
+      }
+    ],
+    "description": "Direction driving the positive monitor statistic. Null means unavailable or not applicable, never zero."
+  },
+  "effect": {
+    "additionalProperties": false,
+    "description": "Latest reading compared with the current post-reset reference.",
+    "properties": {
+      "delta": {
+        "anyOf": [
+          {
+            "description": "Latest reading minus reference median, in input units.",
+            "type": "number"
+          },
+          {
+            "type": "null"
+          }
+        ],
+        "description": "Latest reading minus reference median, in input units. Null means unavailable or not applicable, never zero."
+      },
+      "direction": {
+        "description": "Direction relative to the preceding reference median.",
+        "enum": [
+          "up",
+          "down",
+          "flat"
+        ],
+        "type": "string"
+      },
+      "percentile": {
+        "anyOf": [
+          {
+            "description": "Percentage of reference readings strictly below the current value.",
+            "type": "number"
+          },
+          {
+            "type": "null"
+          }
+        ],
+        "description": "Percentage of reference readings strictly below the current value. Null means unavailable or not applicable, never zero."
+      },
+      "robust_z": {
+        "anyOf": [
+          {
+            "description": "Departure divided by 1.4826 times reference MAD; unavailable at zero spread.",
+            "type": "number"
+          },
+          {
+            "type": "null"
+          }
+        ],
+        "description": "Departure divided by 1.4826 times reference MAD; unavailable at zero spread. Null means unavailable or not applicable, never zero."
+      }
+    },
+    "required": [
+      "direction",
+      "delta",
+      "robust_z",
+      "percentile"
+    ],
+    "type": "object"
+  },
+  "effect_reference_size": {
+    "description": "Preceding observations in the current reference.",
+    "minimum": 0,
+    "type": "integer"
+  },
+  "error": {
+    "description": "Validation or computation failure message; inspect MCP isError or the HTTP status.",
+    "type": "string"
+  },
+  "evidence": {
+    "description": "Compatibility alias of monitoring_statistic; not an e-value.",
+    "type": "number"
+  },
+  "guarantee": {
+    "description": "No statistical or market-performance guarantee is made.",
+    "type": "null"
+  },
+  "interpretation": {
+    "description": "Limits on interpreting a threshold crossing.",
+    "type": "string"
+  },
+  "is_e_value": {
+    "const": false,
+    "description": "The statistic is not a calibrated e-value.",
+    "type": "boolean"
+  },
+  "method": {
+    "description": "Monitor construction and one-sided/two-sided mode.",
+    "type": "string"
+  },
+  "monitoring_statistic": {
+    "description": "Descriptive Shiryaev-Roberts statistic.",
+    "type": "number"
+  },
+  "n": {
+    "description": "Number of supplied observations.",
+    "minimum": 0,
+    "type": "integer"
+  },
+  "reading": {
+    "description": "Interpretation, including remaining warmup requirements.",
+    "type": "string"
+  },
+  "readings_since_change": {
+    "anyOf": [
+      {
+        "description": "Observations after the most recent alarm.",
+        "minimum": 0,
+        "type": "integer"
+      },
+      {
+        "type": "null"
+      }
+    ],
+    "description": "Observations after the most recent alarm. Null means unavailable or not applicable, never zero."
+  },
+  "state": {
+    "description": "Current descriptive monitor state.",
+    "enum": [
+      "warming_up",
+      "steady",
+      "watch",
+      "changed"
+    ],
+    "type": "string"
+  },
+  "statistic_type": {
+    "const": "shiryaev_roberts",
+    "description": "Monitor family.",
+    "type": "string"
+  },
+  "validity": {
+    "const": "descriptive",
+    "description": "Descriptive assessment, not a probability or statistical certificate.",
+    "type": "string"
+  },
+  "watch_at": {
+    "description": "Descriptive watch threshold.",
+    "type": "number"
+  },
+  "weighted": {
+    "description": "Whether half-life weighting was requested.",
+    "type": "boolean"
+  }
+}

First observedOct 8, 2026

TDQS

A4.2/5.0

Behavior3/5

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

Annotations already declare this as read-only, idempotent, non-destructive, so the safety profile is covered. The description adds genuinely useful behavior beyond that: an eight-reading warmup, alarm-triggered reference reset, the caveat that the evidence field is a compatibility alias rather than a calibrated e-value, and that no alarm does not establish stability.

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

Conciseness4/5

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

Front-loads the core purpose and the routing guidance before the mechanics and return values. Dense but each sentence carries load; only the enumerated output-field dump and the JSON example feel slightly padded.

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

Completeness5/5

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

Complete given the complexity: routing, warmup semantics, state values, output fields, a caveat about the evidence alias, and a runnable example. Even though an output schema exists, the prose explains semantics an agent needs to interpret the result correctly.

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

Parameters3/5

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

Schema description coverage is 100%, so the schema already documents values, half_life, and two_sided thresholds, matching the baseline of 3. The description restates the warmup/threshold behavior in prose but adds no parameter syntax or format detail beyond what the schema already provides.

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

Purpose5/5

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

Names a specific verb+resource (scan a chronological numeric metric for departures) and distinguishes itself from siblings by naming real_or_sampling, market_assessment, and spectral_assessment with the conditions that select each. An agent can route between tools without opening a schema.

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

Usage Guidelines5/5

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

Explicitly states when to use this tool (change-review queue) and, critically, when to use each alternative instead. It also adds a negative usage rule: do not use the evidence field to assert valid_evalues in which_metrics_matter.

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

All NoiseFloor — market attention and evidence checks tools Inspect

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.