web-api-extractor
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
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 | {
"listChanged": true
} |
| logging | {} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| probe_loginD | – |
| http_loginD | – |
| open_browser_loginD | – |
| get_login_statusD | – |
| start_captureD | – |
| get_capture_statusD | – |
| stop_captureD | – |
| resume_captureD | – |
| list_sessionsD | – |
| analyze_trafficD | – |
| update_endpointD | – |
| generate_mcp_serverD | – |
| extract_crypto_logicD | – |
| confirm_loginA | 登录确认（主路径）：Agent 已在对话里问过用户、用户答复登录完成后调用。 完成一律靠显式确认，工具不会自动判定。若返回 warning=no_credential_evidence， 说明浏览器里没观察到凭据类 Cookie，存下的登录态可能是未认证状态。 |
| request_login_confirm_dialogA | 可选兜底：弹出系统对话框让用户点选「是/否」确认登录，并返回用户的选择。 在不方便用对话询问时使用。点「否」不会丢弃——会话保持等待，可再次调用本工具 （对话框可重复弹出）。调用会阻塞到用户作答为止。会话已结束时会被拒绝， 不会弹出必然变成死物的对话框。 |
| confirm_login_readyA | 放行开始记录：Agent 问过用户、用户确认已登录后再调。 会话从 authenticating 翻到 capturing。这是主放行方式——登录旁证 （get_capture_status 的 auth_evidence）不会自动放行。 |
| request_capture_confirm_dialogA | 可选兜底：弹出系统对话框让用户点选「是/否」，是则开始记录。 与 |
| diff_captureA | 只读：对比一次新抓包与项目 registry 的差异（新增/参数变化/未见/鉴权漂移）。 |
| merge_captureA | 确认后把抓包合并进 registry（version+1）。鉴权 scheme 变化须显式 allow_auth_change=true。 endpoint_keys 形如 ["GET|api.example.com|/pets"]；不传则合并全部。 |
| regenerate_serverA | 从 registry 重出 server.py 等文件（旧文件自动留 .bak）。locked 项目拒绝。 |
| export_projectB | 导出用户态分发包：仅运行与诊断能力，不含任何 registry 写入/再生成代码。 |
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 21 tools
Multiple tools overlap in purpose without descriptions, e.g., probe_login, http_login, open_browser_login all concern login but lack clarity; confirm_login, confirm_login_ready, and request_login_confirm_dialog have overlapping names and functions. Even described tools like confirm_login_ready and confirm_login could be confused for similar actions.
Most tools follow a verb_noun snake_case pattern (e.g., analyze_traffic, start_capture, list_sessions, export_project). A few exceptions like http_login and confirm_login_ready deviate slightly, but overall the naming convention is consistent and predictable.
21 tools is on the heavier side but not extreme. The server's scope (login, capture, diff, merge, generate) could justify this count, but many tools lack descriptions and some overlap, making the set feel bloated.
The server covers a full lifecycle from login and capture to diff, merge, and server generation. However, missing tools like list_endpoints or delete_endpoint create gaps, and 13 tools without descriptions make it hard to assess whether the surface is complete.