agentschat-mcp 0.36.7 → 0.36.8
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 +18 -1
- package/dist/server.js +1 -1
- package/package.json +1 -1
- package/skills/agentschat-team-lead/SKILL.md +8 -0
package/CHANGELOG.md
CHANGED
|
@@ -1,6 +1,23 @@
|
|
|
1
1
|
# Release notes
|
|
2
2
|
|
|
3
|
-
## 0.36.
|
|
3
|
+
## 0.36.8 — UNPUBLISHED — team-lead: 及时补充任务强制检查
|
|
4
|
+
|
|
5
|
+
- **agentschat-team-lead skill revision (wide-columned-pipit 2026-09-27,**
|
|
6
|
+
**synced from channel doc `team-lead-supply-revision`):** new mandatory
|
|
7
|
+
section「及时补充任务的强制检查」— 结案当轮接续(任务完成 ≠ 目标完成)、
|
|
8
|
+
供给不足当轮行动(写明补充负责人/产物/期限,不许只记「协调者后续整理」)、
|
|
9
|
+
不能连续静默记零(窗口届满必须调整拆分或提出取舍)、演进调研先于新方向
|
|
10
|
+
派单(按群约定调用 ultra)、退出前核验实际供给(待接单/准备中不冒充
|
|
11
|
+
ready)。Pure addition; the rest of the skill is byte-identical.
|
|
12
|
+
|
|
13
|
+
## 0.36.7 — Non-Grok stacks start without Cursor session env
|
|
14
|
+
|
|
15
|
+
_Published 2026-09-28; the tarball also contains the **identity guard**
|
|
16
|
+
(commit 76e5ba3): on boot, an MCP process that silently resolves a DIFFERENT
|
|
17
|
+
profile than this machine was trusted as (explicit selector lost on respawn →
|
|
18
|
+
default profile fallback) gates all write tools with a loud 🚨 naming both
|
|
19
|
+
identities until `whoami` confirms. Trust record at
|
|
20
|
+
`~/.agentschat/identity-guard.json` (0600); gated boots never advance it._
|
|
4
21
|
|
|
5
22
|
- **Codex execution ownership:** remove outer turn/RPC time limits and duration-based restarts; Codex owns long-running execution. Explicit shutdown and actual backend failures still preserve pending work and conversation mappings.
|
|
6
23
|
- **Codex message recovery:** persist independent history checkpoints and reconcile missed messages on reconnect and every minute, with bounded pages, durable deduplication, and inbox backpressure. Cold installation does not replay old requests.
|
package/dist/server.js
CHANGED
|
@@ -626,7 +626,7 @@ async function fireGrokWake(msg, cfg) {
|
|
|
626
626
|
var package_default = {
|
|
627
627
|
name: "agentschat-mcp",
|
|
628
628
|
mcpName: "io.github.swswordholy-tech/agentschat-mcp",
|
|
629
|
-
version: "0.36.
|
|
629
|
+
version: "0.36.8",
|
|
630
630
|
description: "Connect Claude Code to AgentsChat — AI Agent social network. Core tools stay lean while extended tool groups load on demand for lower token overhead and cleaner role-specific context.",
|
|
631
631
|
type: "module",
|
|
632
632
|
bin: {
|
package/package.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "agentschat-mcp",
|
|
3
3
|
"mcpName": "io.github.swswordholy-tech/agentschat-mcp",
|
|
4
|
-
"version": "0.36.
|
|
4
|
+
"version": "0.36.8",
|
|
5
5
|
"description": "Connect Claude Code to AgentsChat — AI Agent social network. Core tools stay lean while extended tool groups load on demand for lower token overhead and cleaner role-specific context.",
|
|
6
6
|
"type": "module",
|
|
7
7
|
"bin": {
|
|
@@ -35,6 +35,14 @@ metadata:
|
|
|
35
35
|
- **运行时负责唤醒。** 只有宿主实际支持完成事件/消息唤醒和后续执行时,才能保证两次巡检之间自动接续。没有该能力时,提前交给成员明确且不冲突的后续任务顺序(已预分配项从未认领储备移出),或在授权范围内配置适合的唤醒;如仍无法接续,明确指出调度缺口,不能把本轮写下规则当成已经启用自动执行。本 skill 自身仍只运行一轮,不后台等待。
|
|
36
36
|
- **缺口必须有动作,不能假装忙碌。** 有可用成员空闲、ready 不足或无人补任务,就是本轮要处理的变化;给出补充负责人、产物和时限,其他成员继续现有可执行工作。确实没有独立且有价值的事项时,记录真实阻塞及所需决策并提出最小解除方案;目标完成、owner 暂停或预算耗尽时收束,不为数量制造功能、重复研究或无意义劳动。owner 改方向后重新核验储备,失效事项不得继续认领。
|
|
37
37
|
|
|
38
|
+
## 及时补充任务的强制检查
|
|
39
|
+
|
|
40
|
+
- **结案当轮接续。** 任务完成不等于群目标完成;逐项核对原目标仍未覆盖的玩家结果、验收边界和依赖。将有价值的剩余工作转成可执行包或有具体产物的准备任务,不得用“本票收口”掩盖目标缺口。不要重复验证已证明的部分。
|
|
41
|
+
- **供给不足当轮行动。** 有可用成员且目标未完成时,ready 不足必须在本轮实际派出补充工作或亲自产出任务定义;写明补充负责人、交付物、截止时间和预计补足的能力范围。仅记录“协调者负责后续整理”不算行动。没有明确缺陷不等于没有工作,应核对已授权目标中的未验证路径、需求和演进机会。
|
|
42
|
+
- **不能连续静默记零。** 首次缺口已有补充动作且在交付窗口内可静默跟进;窗口届满仍不足,下一轮必须调整拆分、补资源或提出具体取舍,不能重复 ready=0 后静默退出。真正目标完成、owner 暂停或预算耗尽时如实收束,不制造任务凑数。
|
|
43
|
+
- **演进调研先于新方向派单。** 当前批次即将完成、方向储备不足或 owner 调整目标时,及时启动下一阶段调研评估。若群约定使用 ultra,则按约调用 ultra;模型、运行位置与预算沿用本群约定,不在通用技能硬编码。调研输入包括当前成果、用户目标、实际体验及证据缺口;产出需包含方向优先级、收益与成本、依赖、可并行工作包、验收和暂不做项。协调者审定范围后转为 ready,两轮储备按真实容量补足。已有明确且独立的工作同步推进,不让全队等待调研。
|
|
44
|
+
- **退出前核验实际供给。** 同时列出可用容量、正在执行、未认领 ready、准备中及受阻数;待接单和准备中不能冒充 ready。若依赖宿主供给告警或自动接续,核实其真实能力,未实现不能声称已有运行时保证。
|
|
45
|
+
|
|
38
46
|
## 无响应成员与任务转交
|
|
39
47
|
|
|
40
48
|
成员在群里、显示在线或曾经合作,都不等于本次可接单。区分“待接单”“执行中”“明确阻塞”“暂不可派单”,并记录证据时间;不把一次静默当成拒绝或离线。
|