Skip to main content
Glama
Sign In
Kiln

Kiln

by codeofaxel
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python

record_calibration_correction

Destructive

Record a calibration correction event that links a wrong observation to an optional correcting observation, enabling trust-weight adjustments and an audit trail.

Instructions

Record an explicit correction event linking two observations.

Needs Kiln Pro.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "original_observation_id": "<original_observation_id>"
    },
    "name": "record_calibration_correction"
  }
}

Input Schema

NameRequiredDescriptionDefault
reasonNoOne of ``"user_explicit"`` (the user said so), ``"agent_review"`` (an agent flagged it), ``"opposite_sign_significant_delta"`` (auto-flagged by the heuristic detector), or ``"auto_flagged_at_ingest"`` (surfaced at ingest time and now confirmed). Default: "user_explicit".
annotationNoFree-form context. Default: "".
confidence_adjustmentNoNegative-or-zero float in ``[-1, 0]`` — how much trust the corresponding parsing rule should lose. ``0.0`` = audit only; ``-0.5`` = halve future weights. Default: 0.0.
original_observation_idYesThe observation that was wrong.
corrected_by_observation_idNoOptional — the observation that proves it wrong (typically a later caliper measurement). Empty string means "user just told us it was wrong, no specific corrector". Default: "".

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed5 schema fields changedv1.5.1
    • changedInput schema / properties / annotation / description
      Previous value: -"Default: \"\"."New value: +"Free-form context. Default: \"\"."
    • changedInput schema / properties / confidence_adjustment / description
      Previous value: -"Default: 0.0."New value: +"Negative-or-zero float in ``[-1, 0]`` — how much trust the corresponding parsing rule should lose.  ``0.0`` = audit only; ``-0.5`` = halve future weights. Default: 0.0."
    • changedInput schema / properties / corrected_by_observation_id / description
      Previous value: -"Default: \"\"."New value: +"Optional — the observation that proves it wrong (typically a later caliper measurement).  Empty string means \"user just told us it was wrong, no specific corrector\". Default: \"\"."
    • addedInput schema / properties / original_observation_id / description
      Added value: +"The observation that was wrong."
    • changedInput schema / properties / reason / description
      Previous value: -"Default: \"user_explicit\"."New value: +"One of ``\"user_explicit\"`` (the user said so), ``\"agent_review\"`` (an agent flagged it), ``\"opposite_sign_significant_delta\"`` (auto-flagged by the heuristic detector), or ``\"auto_flagged_at_ingest\"`` (surfaced at ingest time and now confirmed). Default: \"user_explicit\"."
  2. Addedv1.5.0

TDQS

B3.2/5.0
Behavior3/5

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

Annotations already supply destructiveHint=true and readOnlyHint=false, and the description adds genuinely useful plan-gating context ('Needs Kiln Pro') that annotations cannot express. Beyond that it says nothing about irreversibility, what happens to the original observation's standing, or whether the event persists as an audit record only — notable gaps for a destructive write.

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?

Two short sentences, front-loaded with the action and followed by the prerequisite — no filler. The 'Needs Kiln Pro' fragment is terse to the point of being slightly cryptic about what tier gating actually implies.

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

Completeness3/5

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

With five parameters, a destructive annotation, and no output schema, the definition should say more about consequences and the resulting record. It is adequate to select and call the tool correctly, but an agent gets no picture of what the call produces or how it interacts with the other calibration tooling.

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 all five parameters including the reason enum values and the confidence_adjustment range. The phrase 'linking two observations' mildly clarifies the relationship between original_observation_id and the optional corrected_by_observation_id, but adds no syntax beyond that.

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

Purpose4/5

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

States a specific verb and resource: 'Record an explicit correction event linking two observations.' The linking of two observations is a concrete, distinguishing detail. It falls short of 5 because it never distinguishes itself from calibration siblings such as revert_calibration_recommendations or ingest_measurement.

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

Usage Guidelines2/5

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

The only guidance is the gating prerequisite 'Needs Kiln Pro.' There is no statement of when to record a correction versus reverting recommendations, ingesting a measurement, or applying recommendations — an agent must infer the trigger conditions entirely.

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

Deploy Server

Other Tools