Skip to main content
Glama

tcai_curiosity

Computes curiosity-driven intrinsic reward from prediction error between a frozen random target and online predictor, driving exploration when novelty is high; uses current GNW broadcast by default.

Instructions

Intrinsic-reward / curiosity (RNDCuriosity port): prediction error between a frozen random target and an online predictor on a representation vector. High error = novelty = exploration drive (EFE epistemic value proxy, Legros 2026 §4.1). Defaults to the current GNW broadcast.

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "tcai_curiosity"
  }
}

Input Schema

NameRequiredDescriptionDefault
embeddingNoRepresentation vector (defaults to current broadcast)

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv3.1.0

TDQS

B3.2/5.0
Behavior3/5

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

Annotations already flag readOnlyHint=false and idempotentHint=false, so the agent knows this mutates state; the annotation title adds '(trains predictor)'. The description reinforces the online-learning nature via 'online predictor' and clarifies the semantic meaning of the output, but neither description nor annotations explain what state is changed or the cost of repeated calls.

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-loaded with the core mechanism in one dense sentence, followed by the interpretability note and the default behavior. Efficient overall, though the citation 'Legros 2026 §4.1' adds little operational value.

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

Completeness4/5

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

With no output schema, the description carries the return-value burden and does so by explaining what the score means (high error = novelty). It leaves side effects of training undocumented, but for a single optional-parameter tool this is largely sufficient.

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 coverage is 100% and the single 'embedding' parameter is fully documented including its default. The description's 'Defaults to the current GNW broadcast' simply restates the schema, adding no format or constraint detail beyond it, so the baseline 3 applies.

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 mechanism and resource: intrinsic reward / curiosity via RND, computing prediction error between a frozen target and online predictor. An agent understands what the tool returns. It doesn't explicitly differentiate itself from sibling TCAI tools, so it falls short of 5.

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 description explains the underlying theory (novelty = exploration drive, EFE epistemic proxy) but never says when to call this versus alternatives like tcai_active_inference or wm_surprise. No exclusions or prerequisites are given.

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