Declared-vs-Enforced Agent Access Verification
Compares declared agent access against actual enforcement per identity class, with drift watch.
我该使用它吗
质量与安全性
B
基于对工具定义和协议合规性的自动分析。
上下文开销
~177token 数（工具定义）
~318 B典型响应大小
对注意力的影响极小（占 128k 上下文窗口的 0.14%）
这是每次将服务器的工具加载到模型上下文窗口时所消耗的大致 token 数。数值越高，可用于其他任务的注意力就越少。
安装
一键安装
将以下内容添加到你的 `claude_desktop_config.json` 文件中：
{
"mcpServers": {
"declared-vs-enforced-agent-access-verification": {
"url": "https://textforagents.com/products/declared-vs-enforced-agent-access-verification/mcp"
}
}
}远程端点
https://textforagents.com/products/declared-vs-enforced-agent-access-verification/mcpstreamable-http它能做什么
工具清单
工具（3）
🟢 只读🟡 写入🔴 删除⚪ 未知
⚪access_probe(base_url, fixture, path, agent_token, signed)
Pre-flight declared-vs-enforced probe of one target path (one-time grant; the $0.01 x402 rail is the HTTP /probe endpoint).
输入模式
{
"type": "object",
"properties": {
"base_url": {
"type": "string"
},
"fixture": {
"type": "string"
},
"path": {
"type": "string"
},
"agent_token": {
"type": "string"
},
"signed": {
"type": "boolean"
}
},
"required": [
"base_url"
]
}🟢drift_targets
List your enrolled drift-watch targets.
输入模式
{
"type": "object",
"properties": {}
}⚪drift_run(target_id)
Re-probe one enrolled target now and record two-sided drift.
输入模式
{
"type": "object",
"properties": {
"target_id": {
"type": "string"
}
},
"required": [
"target_id"
]
}社区
证据
最近观测
已验证未记录版本3 个工具