Unverified50% confidenceOpinionExact time
When the Agent that wrote the code reviews it afterward, it compares the code with the intent rather than the code with the result, so it easily concludes that it is "done".
1
Sources
50%
Confidence
Long-term
Relevance
10/19/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedCode is inherently executable and testable, and the runtime result itself is an objective criterion for correctness, requiring no subjective scoring.74% similarVerifiedExplanation tasks can use existing code as an anchor for semantic analysis, and the correctness of their output can be constrained and verified by the code itself.73% similarUnverified该工作流采用生成→审查→修补→再审查的循环迭代，被标记的问题由Agent自动打补丁，直到代码干净通过测试73% similarUnverifiedThe Inner Check works by appending a mandatory self-verification statement at the end of a skill, shifting the Agent's evaluation criterion from execution intent to actual output.73% similarUnverifiedIt is suggested to perform end-to-end verification of write operations, not relying on the API's success response, explicitly verifying in subsequent steps whether the write took effect, and recording the verification result in the trace.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1066106API
curl https://kongchang.com/api/v1/knowledge/claims/1066106MCP
get_claim(id=1066106)