Search explanations describe parts the search didn't run
維護者通常 1 天內回覆
還沒有人認領這個 Issue。
評估
- 難度
- 4/5
- 預估耗時
- 3-5 天
- 新手友好度
- 57/100
- Issue 類型
- 缺陷
- 描述清晰度
- 描述清楚
- 活躍度
- 活躍
- 技術堆疊
- typescript
研究方向
Start with the time-series agent output in agent.ts:53-72, then follow how search.ts:910-948 builds the executed request and how structured.ts:462-469 and formatters.ts:209-210 return the explanation. Compare the explanation with the request that actually ran, including fields and time range; done means callers are not given an explanation that claims omitted search parts were executed.
由索引模型根據 Issue 內容生成。
描述
includeExplanation on search_errors and the other runSearchEvents tools returns the embedded agent's own explanation text unchanged. It can describe grouping, fields, or a time range that the executed request does not contain, so callers trust results that don't match what they asked for.
Reported in Sentry MCP user feedback (MCP-SERVER-GN4) from an automated caller:
- Grouped time series: an hourly series asked for grouping by
projectplus distinct users. The explanation said both were returned.executedSearch.fieldswas onlycount(), and the output was one ungrouped series. The span for that call was not sampled, so this case is from the feedback text only. - Time range: the explanation described a fixed window while
executedSearch.timeRangewasstatsPeriodwith nullstart/end. This one is verified in a trace and fixed in #2146.
What the code does (main @ 737f183):
- The agent's time-series output is one
yAxisplusinterval, with no group-by or top-N field (agent.ts:53-72). The handler sends oneyAxistoevents-timeseries(search.ts:1083-1104) and reportsfields: [yAxis](structured.ts:469). Any grouping or second aggregate the agent planned is dropped without a note. explanationis set fromparsed.explanation(search.ts:932) and passed through unchanged (structured.ts:462-465,formatters.ts:209-210). The handler later overrides the time range, fields, sort, or query (search.ts:910-948), but never compares the explanation with what ran.
Related: #2017 adds explicit multi-axis series for traces, logs, and metrics, not errors and not grouped series. #2139 and #2146 fix time-range drops, not the explanation text.
via Sentry.
--
- 主要語言
- TypeScript
- 星號
- 961
- 分支
- 179
- 平均合併
- 15 小時 33 分鐘
- 30 天內合併 PR
- 476
環境準備
這個專案沒有提供開發容器、Dockerfile 或貢獻指南,環境需要你自己搭建:先看它的 README,通用步驟見我們的新手貢獻指南。
從這裡開始
- 先讀完整個 Issue,再讀專案的貢獻指南。
- 在 Issue 下留言說明你要接手 —— 這能避免兩個人做同樣的事。
- Fork 儲存庫,在一個分支上完成修改。
- 送出 Pull Request,並在描述裡引用這個 Issue 編號。
getsentry/toolkit 的其他 Issue
-
MCP mcp-feedback-pattern
難度 2/5 1-3 小時 新手友好度 78/100
getsentry/toolkit#2276 · 1 則留言 ·
維護者通常 1 天內回覆
-
CLI enhancement MCP
難度 2/5 1-3 小時 新手友好度 63/100
getsentry/toolkit#1908 · 1 則留言 ·
維護者通常 1 天內回覆
-
難度 2/5 1-3 小時 新手友好度 72/100
getsentry/toolkit#1471 · 1 則留言 ·
維護者通常 1 天內回覆
-
get_sentry_resource gives no guidance for Explore > Errors list URLs可能已有人在做 @sentry-junior 今天認領。 未關閉MCP mcp-feedback-pattern
難度 2/5 1-3 小時 新手友好度 32/100
getsentry/toolkit#2278 · 1 則留言 ·
維護者通常 1 天內回覆
-
難度 3/5 1-2 天 新手友好度 65/100
getsentry/toolkit#2275 · 1 則留言 ·
維護者通常 1 天內回覆
查看 getsentry/toolkit 的全部 Issue
相似的 Issue
-
難度 2/5 1-3 小時 新手友好度 72/100
AykutSarac/jsoncrack.com#616 ·
-
nip27: long hashtag swallows the rest of the content, lone # gives an empty hashtag可能已有人在做 @bardrop 今天認領。 未關閉
難度 2/5 1-3 小時 新手友好度 82/100
nbd-wtf/nostr-tools#578 ·
-
難度 2/5 1-3 小時 新手友好度 68/100
backnotprop/plannotator#3427 ·
維護者通常 1 天內回覆
-
bug
難度 2/5 1-3 小時 新手友好度 72/100
microsoft/Power-CAT-Copilot-Studio-Kit#942 ·
維護者通常 1 天內回覆
-
[BUG]: Duplicate preposition in heading and unformatted feedback text on Strengthen Argument page未關閉bug
難度 2/5 1-3 小時 新手友好度 72/100
AOSSIE-Org/DebateAI#801 · 1 則留言 ·
維護者通常 1 天內回覆