steward
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| SHADOWROOM_STEWARD_DB | No | Overrides the ledger database path. Default is ~/Documents/ShadowRoom/_steward/ledger.db |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| report_doneC | 一个工作区干完活后的回执(谁、什么、结果、文件、≤3 个数字)。 |
| recent_eventsC | 按 id 拉事件明细(管家要细节时才用)。 |
| situation_reportC | 随时汇报:最近关键事件、未解决的问题、活跃的工作区。 |
| focus_changeC | 记录焦点变化(切页面/切工程/选中片段)。 |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 4 tools
report_done creates a completion receipt, situation_report summarizes current status, recent_events fetches event details by id, and focus_change records a focus shift. The purposes are mostly distinct, though report_done and situation_report both involve reporting and could be confused in edge cases.
All names use snake_case and are readable, but they follow mixed patterns: report_done is verb-participle, recent_events is adjective-noun, while situation_report and focus_change are noun-noun. There is no single predictable verb_noun convention.
Four tools is a reasonable, well-scoped set for a lightweight steward assistant. Each tool appears to earn its place, though the surface is minimal enough that it may feel slightly thin for full workspace tracking.
The set covers reporting completion, summarizing status, retrieving event details, and recording focus changes. However, it lacks explicit workspace lifecycle operations and event discovery beyond lookup by known id, which may cause dead ends for agents.