Unverified50% confidenceOpinionExact time
Most bizarre agent behaviors are rooted in poor tool descriptions, because the model guesses meaning from descriptions rather than reading code logic.
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe initial results of an agent generated from a single sentence are often imperfect, making it suitable for quickly validating ideas and handling simple requirements75% similarUnverified模糊的Skill描述会导致模型错误认为某Skill适用于当前场景，这一现象被称为工具幻觉（Tool Hallucination）75% similarUnverifiedVague tool authorization prompt descriptions can easily trigger completely different execution strategies in different models.74% similarVerifiedWhen a model lacks accurate knowledge of a particular library, it tends to generate syntactically correct but semantically incorrect code—so-called 'confident errors.'73% similarVerifiedResearch indicates that models may produce unfaithful chains of thought, where the surface-level reasoning steps are inconsistent with the actual basis for decision-making.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/977130API
curl https://kongchang.com/api/v1/knowledge/claims/977130MCP
get_claim(id=977130)