Unverified50% confidenceDecision RuleExact time
When designing tool-call logs, wrong tool selection and parameter errors should be recorded in separate tables; otherwise, subsequent attribution analysis cannot distinguish routing problems from argument-passing problems.
1
Sources
50%
Confidence
Long-term
Relevance
10/20/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedError classification is a key step in the recovery process, and different error types point to different strategies: timeouts may be retryable, retrying authentication failures is useless, and permission errors require a different path.77% similarVerifiedThe key to error analysis is pinpointing exactly which stage and which type of input caused the failure, rather than blindly adjusting prompts.76% similarUnverifiedCommon tool-calling evaluation frameworks break tool failures into two major categories: "wrong tool selected" and "correct tool selected but wrong parameters," with parameter errors further divided into structure and value validation.76% similarUnverifiedSchema validation for tool calls can only catch structural issues (type errors, missing required fields) and cannot catch semantic errors.76% similarUnverifiedTo track and reproduce failures, each evaluation case needs to record the question, expected answer, retrieved document IDs and chunks, similarity scores, prompt and model configuration, generated answer, metric results, latency, and token usage, while avoiding recording keys or sensitive user data.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1074126API
curl https://kongchang.com/api/v1/knowledge/claims/1074126MCP
get_claim(id=1074126)