laminar-mcp

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Laminar MCP

Laminar MCP

Overview

概述

Laminar is the product behind this MCP. Server
tools/list
(or equivalent)
wins on names and args over this file. No assumption of a local Laminar checkout. Prefer signals from
load_source_context
; stop loading when the next step is clear.
Laminar 是这款MCP背后的产品。服务器端
tools/list
(或等效工具)
在命名和参数方面的优先级高于本文件。无需假设本地存在Laminar代码仓库。优先使用
load_source_context
返回的信号;当下一步操作明确时停止加载。

For the developer (human)

面向开发者(人工)

  • Attach
    @laminar-mcp
    when starting demand work so the agent reads this playbook.
  • Give demand id (
    TAL-*
    or Convex id) and client/product if scope must not be guessed; say if you want anchored handoff at session end.
  • If MCP fails: confirm Laminar MCP enabled + auth; retry or paste
    tools/list
    for the agent.
  • 开始处理需求时添加**
    @laminar-mcp
    **,以便Agent读取本操作手册。
  • 若范围不能被推测,请提供需求ID
    TAL-*
    或Convex ID)和客户端/产品信息;若希望在会话结束时进行锚定交接,请说明。
  • 若MCP故障:确认Laminar MCP已启用且认证通过;重试或为Agent粘贴**
    tools/list
    **的结果。

Agent: session scope

Agent:会话范围

get_current_context
→ if needed
list_clients
/
list_products
set_context
(
clientId
, then
productId
). Lists need client; story map and releases need product.
clear_context
only to drop product on purpose.
get_current_context
→ 若需要则调用
list_clients
/
list_products
set_context
(传入
clientId
,再传入
productId
)。列表查询需要客户端信息;故事图和版本查询需要产品信息。仅当有意移除产品信息时调用
clear_context

Agent: demand context (order)

Agent:需求上下文(顺序)

Demand record → anchored ADR → evidence → map → peers.
GoalToolNotes
Demand fields
get_work_item
query
= id.
descriptionFormat: "plain"
to read;
json
/
both
only if editing description.
Handoff doc
get_work_item_anchored_context
Version + markdown.
Pick docs
list_source_contexts
IDs + short lines only—not a planning basis.
Evidence
load_source_context
Per relevant id before substantive answers. Signals default;
includeRawText: true
only if signals missing or insufficient.
Map
get_story_map
Then
list_story_map_activities
/
list_story_map_steps
(
storyMapId
),
list_story_map_work_items
,
list_releases
,
list_work_items
.
Synthesize intent from
get_work_item
+
get_work_item_anchored_context
; ground details in loaded source contexts.
需求记录 → 锚定ADR → 证据 → 故事图 → 关联项。
目标工具说明
获取需求字段
get_work_item
query
参数为需求ID。使用
descriptionFormat: "plain"
来读取内容;仅在编辑描述时使用
json
/
both
格式。
获取交接文档
get_work_item_anchored_context
返回版本信息+Markdown格式内容。
选择文档
list_source_contexts
仅返回ID+简短描述——不可作为规划依据。
获取证据
load_source_context
在给出实质性回答前,按相关ID逐一加载。默认使用信号;仅当信号缺失或不足时设置
includeRawText: true
获取故事图
get_story_map
之后可调用
list_story_map_activities
/
list_story_map_steps
(传入
storyMapId
)、
list_story_map_work_items
list_releases
list_work_items
get_work_item
+
get_work_item_anchored_context
中综合意图;基于已加载的源上下文补充细节。

Agent: handoff

Agent:交接流程

Use
put_work_item_anchored_context
:
get_work_item_anchored_context
→ merge sections per live server schema (
decisions
,
constraints
,
openQuestions
,
stateItems
,
outOfScope
,
scenarios
+ coverage enum, optional
documentTitle
) → put with
expectedVersion
(
0
if new) → on conflict, refetch and retry. Optional:
update_work_item
via TipTap JSON from
get_work_item
(
descriptionFormat: "json"
); preview with
confirmed: false
on writes (
update_work_item
,
transition_work_item
,
assign_work_item
,
unassign_work_item
).
使用**
put_work_item_anchored_context
:调用
get_work_item_anchored_context
→ 根据
服务器实时 schema**(
decisions
constraints
openQuestions
stateItems
outOfScope
scenarios
及覆盖枚举,可选
documentTitle
)合并章节 → 传入
expectedVersion
(若为新文档则传
0
)进行提交;若发生冲突,重新获取后重试。可选操作:通过
get_work_item
返回的TipTap JSON(
descriptionFormat: "json"
)调用
update_work_item
;执行写入操作(
update_work_item
transition_work_item
assign_work_item
unassign_work_item
)时,设置
confirmed: false
来预览结果。

Agent: story-map peers

Agent:故事图关联项

With product in
set_context
:
get_story_map
list_story_map_work_items
→ find row → filter others by same
userStepId
,
userActivityName
, or
releaseId
/
releaseName
→ selectively
get_work_item
/ anchored /
load_source_context
.
set_context
中设置产品信息后:调用
get_story_map
list_story_map_work_items
→ 找到对应行 → 按相同的
userStepId
userActivityName
releaseId
/
releaseName
筛选其他项 → 选择性调用
get_work_item
/ 获取锚定上下文 /
load_source_context

Supporting reads

辅助查询

list_statuses
,
list_team_members
,
list_blocker_types
. Prefer reads first;
confirmed: false
to preview mutations.
list_statuses
list_team_members
list_blocker_types
。优先执行查询操作;使用
confirmed: false
预览变更操作。

Common mistakes

常见问题

SymptomFix
Plans from list lines only
load_source_context
each id.
Raw dumps by defaultSignals first; raw only if insufficient.
Story map / releases errorsSet
clientId
+
productId
.
Chat-only “done”
put_work_item_anchored_context
+ version.
Tool errorsHuman fixes MCP or pastes
tools/list
; no invented args.
症状解决方法
仅基于列表行生成计划按每个ID调用
load_source_context
默认返回原始数据优先使用信号;仅当信号不足时返回原始数据。
故事图/版本查询报错设置
clientId
+
productId
仅通过聊天标记“完成”调用
put_work_item_anchored_context
并传入版本号。
工具报错人工修复MCP或粘贴
tools/list
结果;请勿自行编造参数。

When MCP is down or unclear

当MCP故障或不明确时

Ask the human to restore connectivity or paste
tools/list
. Do not invent parameters.
请用户恢复连接或粘贴**
tools/list
**的结果。请勿自行编造参数。