@lark-project/meegle 1.0.23 → 1.0.25
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +17 -0
- package/README.md +22 -1
- package/README.zh-CN.md +22 -1
- package/bin/meegle-darwin-arm64 +0 -0
- package/bin/meegle-darwin-x64 +0 -0
- package/bin/meegle-linux-arm64 +0 -0
- package/bin/meegle-linux-x64 +0 -0
- package/bin/meegle-win32-arm64.exe +0 -0
- package/bin/meegle-win32-x64.exe +0 -0
- package/package.json +1 -1
package/CHANGELOG.md
CHANGED
|
@@ -8,6 +8,23 @@ versioned section on each npm release.
|
|
|
8
8
|
|
|
9
9
|
## [Unreleased]
|
|
10
10
|
|
|
11
|
+
## [v1.0.25] - 2026-10-13
|
|
12
|
+
|
|
13
|
+
### Added
|
|
14
|
+
|
|
15
|
+
- Added sampled reporting of failed CLI commands after a one-time interactive notice, including command names, failure stages and error codes. Reporting and policy refresh run in the background. Local `meegle telemetry enable|disable` controls let users turn collection on or permanently off; disabling clears queued diagnostics. Argument values, credentials, file paths, payload bodies and error messages are excluded.
|
|
16
|
+
|
|
17
|
+
### Fixed
|
|
18
|
+
|
|
19
|
+
- Linux release binaries are statically linked and do not require a system C library.
|
|
20
|
+
|
|
21
|
+
## [v1.0.24] - 2026-10-08
|
|
22
|
+
|
|
23
|
+
### Changed
|
|
24
|
+
|
|
25
|
+
- Updated the Meegle Agent Skill with confirmed `free_add` option creation for select and multi-select fields, exact-name matching before writes, and unique option ID requirements.
|
|
26
|
+
- Split MQL guidance into focused references and clarified query, pagination, and CLI examples in the Meegle Agent Skill.
|
|
27
|
+
|
|
11
28
|
## [v1.0.23] - 2026-09-08
|
|
12
29
|
|
|
13
30
|
### Changed
|
package/README.md
CHANGED
|
@@ -8,7 +8,7 @@
|
|
|
8
8
|
|
|
9
9
|
Command-line tool for [Meegle](https://meegle.com?utm_source=github&utm_medium=readme&utm_campaign=meegle_cli) ([Lark Project](https://project.feishu.cn?utm_source=github&utm_medium=readme&utm_campaign=meegle_cli)). Manage work items, schedules, and data from your terminal — no browser needed.
|
|
10
10
|
|
|
11
|
-
[Install](#installation) · [Quick Start](#quick-start-human-users) · [Agent Skill](#ai-agent-skill) · [Commands](#commands) · [Enterprise Extensions](#enterprise-cli-extensions) · [Auth](#authentication) · [Config](#configuration) · [Security](#security--risk-warnings) · [Contributing](#contributing)
|
|
11
|
+
[Install](#installation) · [Quick Start](#quick-start-human-users) · [Agent Skill](#ai-agent-skill) · [Commands](#commands) · [Enterprise Extensions](#enterprise-cli-extensions) · [Auth](#authentication) · [Config](#configuration) · [Diagnostics](#failed-command-diagnostics) · [Security](#security--risk-warnings) · [Contributing](#contributing)
|
|
12
12
|
|
|
13
13
|
## Why Meegle CLI?
|
|
14
14
|
|
|
@@ -1030,6 +1030,27 @@ Errors and panic values returned by extension callbacks stay behind a guarded Go
|
|
|
1030
1030
|
Every JSON-RPC response is bounded before decoding: `tools/list` uses an 8 MiB discovery limit and all other calls use a 32 MiB per-response limit.
|
|
1031
1031
|
Dynamic discovery accepts nullable JSON Schema types such as `"type": ["string", "null"]` and exposes them as the underlying CLI/SDK parameter type. Unions containing multiple non-null types are isolated with the stable `unsupported_schema_union` diagnostic instead of silently removing unrelated tools.
|
|
1032
1032
|
|
|
1033
|
+
## Failed-command diagnostics
|
|
1034
|
+
|
|
1035
|
+
After a one-time notice in an interactive terminal, Meegle CLI collects sampled diagnostics for final root commands that fail. Successful commands are not collected in schema v1. Collection never includes command argument values, tokens, file paths, request or response bodies, or error messages.
|
|
1036
|
+
|
|
1037
|
+
The foreground command selects a diagnostic event and hands it to a detached maintenance process. That process queues the event without scanning historical events, then performs cleanup, policy refresh and upload when due. The foreground does not wait for queue/sequence locks, event writes or network delivery. Command return is not confirmation that its event has been queued. Policy is refreshed at most every 12 hours (with a 1-hour retry after an unhealthy refresh). Pending events are isolated by profile and host. Background cleanup removes expired events and the oldest excess events to meet the global limits of 7 days, 1,000 events and 5 MiB. Writes can temporarily exceed these limits until maintenance runs. While the current profile/host has queued events, subsequent commands can trigger global cleanup once a minute, independently of credentials and the one-hour upload retry interval. Recording a new failure during upload cooldown does not retry the network request. Policy refresh also triggers maintenance. Switching to an empty queue does not itself schedule cleanup of inactive profiles. No resident daemon runs after the CLI exits.
|
|
1038
|
+
|
|
1039
|
+
Events, sequence counters and the disclosure marker use temporary files and rename without forcing their data or directories to stable storage. Completed writes remain readable after the writer exits, but power loss or an operating-system crash may lose recent events or roll back counters and the disclosure marker. Counter rollback may affect recovery ordering; a missing disclosure marker requires the notice again before collection. Disclosure writes are independent of the collection switch. Policy and the local collection switch/consent generation retain stable-storage synchronization. Collection is best-effort: worker launch failure, disk errors, prolonged contention or worker termination can lose an event without changing the command's output or exit code. Worker launch and local configuration reads still have a cost; this is not a hard real-time latency guarantee. Delayed events cannot be collected under an earlier local consent generation after disable/re-enable.
|
|
1040
|
+
|
|
1041
|
+
Diagnostics record the failing pipeline stage. A valid report response keeps only the returned failed event IDs queued; an invalid response or failed request retains the whole batch.
|
|
1042
|
+
|
|
1043
|
+
Queues follow the resolved profile and host, including credential-provider account selection. Background delivery requires matching credential provenance (source, provider and authentication header); unmatched or unresolved events remain queued until retention limits remove them. Tokens are not persisted by diagnostics, and stored tokens are not refreshed for background delivery.
|
|
1044
|
+
|
|
1045
|
+
Control collection locally:
|
|
1046
|
+
|
|
1047
|
+
```bash
|
|
1048
|
+
meegle telemetry disable # also clears queued events
|
|
1049
|
+
meegle telemetry enable
|
|
1050
|
+
```
|
|
1051
|
+
|
|
1052
|
+
Set `MEEGLE_TELEMETRY=disabled` to disable it for the current environment. The CLI journey event is separate from MCP server telemetry; when available, a structured backend `logid` is included only as a correlation reference.
|
|
1053
|
+
|
|
1033
1054
|
## FAQ
|
|
1034
1055
|
|
|
1035
1056
|
### Empty Command List
|
package/README.zh-CN.md
CHANGED
|
@@ -8,7 +8,7 @@
|
|
|
8
8
|
|
|
9
9
|
飞书项目([Meegle](https://meegle.com?utm_source=github&utm_medium=readme&utm_campaign=meegle_cli) / [Lark Project](https://project.feishu.cn?utm_source=github&utm_medium=readme&utm_campaign=meegle_cli))命令行工具。在终端中管理工作项、查看排期、搜索数据,无需打开浏览器。
|
|
10
10
|
|
|
11
|
-
[安装](#安装) · [快速开始](#快速开始人类用户) · [Agent Skill](#ai-agent-skill) · [命令](#命令一览) · [企业扩展](#企业-cli-扩展) · [认证](#认证) · [配置](#配置) · [安全](#安全与风险提示) · [贡献](#贡献)
|
|
11
|
+
[安装](#安装) · [快速开始](#快速开始人类用户) · [Agent Skill](#ai-agent-skill) · [命令](#命令一览) · [企业扩展](#企业-cli-扩展) · [认证](#认证) · [配置](#配置) · [诊断](#失败命令诊断) · [安全](#安全与风险提示) · [贡献](#贡献)
|
|
12
12
|
|
|
13
13
|
## 为什么选择 Meegle CLI?
|
|
14
14
|
|
|
@@ -997,6 +997,27 @@ Policy 拒绝遵循显式结构化输出格式,并使用稳定错误码 `CLIEN
|
|
|
997
997
|
所有 JSON-RPC 响应都会在解码前限制大小:`tools/list` 使用 8 MiB 的发现上限,其他调用使用每次响应 32 MiB 的通用上限。
|
|
998
998
|
动态发现支持 `"type": ["string", "null"]` 这类 nullable JSON Schema,并在 CLI/SDK 中按其非 null 类型注册参数。包含多个非 null 类型的 union 会被单项隔离,并输出稳定的 `unsupported_schema_union` 诊断,不影响其他工具。
|
|
999
999
|
|
|
1000
|
+
## 失败命令诊断
|
|
1001
|
+
|
|
1002
|
+
在交互终端展示一次性告知后,Meegle CLI 会按策略采集最终执行失败的根命令诊断;V1 不采集成功命令。采集内容不包含命令参数值、Token、文件路径、请求/响应 Body 或错误消息。
|
|
1003
|
+
|
|
1004
|
+
前台命令筛选诊断事件后,交给独立维护进程。后台不扫描历史事件即可完成单条入队,再按原有条件进行清理、策略刷新与上报。前台不等待队列锁、序号锁、事件写盘或网络投递;命令返回不代表事件已经入队。有效策略最多每 12 小时刷新一次;刷新不健康时每 1 小时重试。待上报事件按 profile + host 隔离,后台按 7 天、1000 条、5 MiB 的全局限制清理过期和最旧事件;维护运行前允许临时超限。当前 profile/host 有积压时,后续命令最多每分钟触发一次全局清理,不依赖凭证,也不受上报一小时重试间隔限制。上报冷却期间仍可记录新错误,但不会因此重试网络请求。策略到期也会触发维护;切到空队列本身不会触发其他 profile 的清理。CLI 退出后没有常驻守护进程。
|
|
1005
|
+
|
|
1006
|
+
事件、序号计数器和告知标记通过临时文件写入并改名保存,不强制将数据及其目录同步到物理磁盘。完成写入的内容在写入进程退出后仍可读取,但断电或操作系统崩溃时,最近事件可能丢失,序号和告知标记可能回退。序号回退可能影响恢复后的队列顺序;告知标记丢失时,必须再次告知才可采集。告知标记独立于采集开关保存;策略、采集开关及用户开关版本仍保留可靠落盘的同步。采集仅尽力完成:后台启动失败、磁盘错误、长时间争用或后台进程被终止时可能漏收,但不改变命令输出和退出码。启动后台进程和读取本地配置仍有开销,不承诺严格的实时耗时上限。关闭再开启后,延迟事件不能沿用旧的用户开关版本入队。
|
|
1007
|
+
|
|
1008
|
+
诊断记录失败所在的流水线阶段。收到合法上报响应后,只保留返回的失败事件 ID;响应非法或请求失败时保留整批。
|
|
1009
|
+
|
|
1010
|
+
队列跟随最终解析出的 profile 和 host,包括凭证扩展选择的账号。后台只投递凭证来源(来源类型、provider 和认证头)匹配的事件;身份未确定或来源不匹配的事件保留到队列保留上限。诊断功能不持久化 Token,后台投递也不刷新存储中的 Token。
|
|
1011
|
+
|
|
1012
|
+
本地开关:
|
|
1013
|
+
|
|
1014
|
+
```bash
|
|
1015
|
+
meegle telemetry disable # 同时清空待上报队列
|
|
1016
|
+
meegle telemetry enable
|
|
1017
|
+
```
|
|
1018
|
+
|
|
1019
|
+
设置 `MEEGLE_TELEMETRY=disabled` 可在当前环境关闭。CLI 端到端事件与 MCP 服务端埋点彼此独立;存在后端 `logid` 时,只把它作为结构化关联字段。
|
|
1020
|
+
|
|
1000
1021
|
## 常见问题
|
|
1001
1022
|
|
|
1002
1023
|
### 命令列表为空
|
package/bin/meegle-darwin-arm64
CHANGED
|
Binary file
|
package/bin/meegle-darwin-x64
CHANGED
|
Binary file
|
package/bin/meegle-linux-arm64
CHANGED
|
Binary file
|
package/bin/meegle-linux-x64
CHANGED
|
Binary file
|
|
Binary file
|
package/bin/meegle-win32-x64.exe
CHANGED
|
Binary file
|