Jul 14, 2026

On-Chain Risk, Funding and Rates, Retrieval Basics, and Agent Runtime

A daily crypto, finance, and AI vocabulary puzzle covering settlement risk, rate concepts, retrieval pipelines, and agent execution controls.

On-Chain Risk

These terms cover cross-chain movement, data dependency, shared signing, and chain replacement events.

BridgeOracleMultisigReorg
  • Bridge: A bridge is a system that transfers assets or messages between blockchains.
  • Oracle: An oracle supplies external data to a smart contract or blockchain application.
  • Multisig: A multisig arrangement requires more than one signature or approval to authorize an action.
  • Reorg: A reorg happens when a blockchain replaces recently accepted blocks with a different chain history.

Funding and Rates

These terms describe interest-style benchmarks, short debt, and repo plumbing in traditional finance.

Basis PointTreasury BillYield CurveRepo Market
  • Basis Point: A basis point is one-hundredth of a percentage point and is commonly used for rates and spreads.
  • Treasury Bill: A treasury bill is a short-term government debt instrument often used as a cash-like benchmark.
  • Yield Curve: A yield curve shows how interest rates differ across bonds with different maturities.
  • Repo Market: The repo market is where participants borrow cash against collateral and later repurchase the collateral.

Retrieval Basics

These terms describe tokens, representations, indexes, and ranking in retrieval-augmented systems.

TokenizerEmbeddingVector StoreReranker
  • Tokenizer: A tokenizer breaks text into tokens that a language model can process.
  • Embedding: An embedding is a numeric representation that captures semantic meaning in a compact form.
  • Vector Store: A vector store holds embeddings so similar items can be retrieved efficiently.
  • Reranker: A reranker reorders retrieved candidates to improve the final result quality.

Agent Runtime

These terms cover tool use, isolation, testing, and saved context in agent workflows.

Tool CallingSandboxEvalsMemory
  • Tool Calling: Tool calling is the act of invoking a function, API, or service from an AI workflow.
  • Sandbox: A sandbox is a controlled environment that limits what code or an agent can affect.
  • Evals: Evals are repeatable tests used to measure whether a model or agent behaves as expected.
  • Memory: Memory is stored context that can help an agent reuse facts or preferences later.