rootandClaude Opus 5.5
c437c6bff6
- browser.screenshot returns the page as an image content block (≤1280px wide, JPEG quality configurable) for vision-capable upstreams; non-vision upstreams get a redacted text snapshot. Login pages / password fields are marked sensitive and withheld unless a browser.sensitive_screenshot grant exists and allowSensitive is passed. - browser.click_at / type_at / scroll / wait act in screenshot pixels; the server converts via viewport/DPR, requires automation control and an allowlisted host, and writes a thumbnail evidence file (DATA_DIR/tmp/evidence) referenced from agent_run_event. - browser.solve_challenge runs a bounded child agent_run (vision caps only, maxSteps ≤ 12, budget), judges success from page facts, stops immediately on 2FA / SMS / device confirmation, and otherwise moves the session to waiting with evidence and a takeover link. - PiAgentRuntime passes modelImage outputs as pi-ai image content when model.input has image; the ledger never stores image data; coordinate actions use per-invocation idempotency keys. - Tests: image codec, runtime image content, E14 e2e (login solved, 2FA stop, step limit, sensitive withheld). Real camofox 1.17.0 check on example.com recorded in the runbook. Co-Authored-By: Claude Opus 5.5 <[email protected]>
@umt/contracts
所有 API、SSE 事件与能力(capability)输入输出的唯一定义(zod 4 schema + 推导出的 TS 类型)。web 与 server 共用,任何一端都不得手写重复 DTO;新增接口先在这里加 schema,再由 server 用 @hono/zod-openapi 生成 OpenAPI。
| 文件 | 内容 |
|---|---|
common.ts |
Id、Timestamp、DateOnly、JsonObject、游标分页、错误码、响应信封、ApiScope、ActorType |
content.ts / identity.ts / assets.ts / sources.ts / policy.ts / execution.ts / agent.ts / ops.ts / organization.ts / instances.ts |
每张表的 Row schema(camelCase、日期为 ISO 字符串)+ XxxCreate / XxxUpdate 输入 |
organization.ts |
另含公开整理记录合同:PublicOrganizationRecord、OrganizationChangesQuery/Response(change_seq 游标)、OrganizeRequest |
settings.ts |
GlobalSettings、DEFAULT_SETTINGS、GlobalSettingsUpdate、mergeSettings() |
capability.ts / capabilities-io.ts |
CapabilityDescriptor、首批 14 个能力 ID 与目录、各能力 Input/Output schema、CAPABILITY_IO |
events.ts |
SSE 信封、EventType、EVENT_PAYLOADS、TypedEvent 判别联合 |
files.ts |
文件管理(/files/*)的请求/响应:根与条目、dryRun 计划(planId)、move/copy/organize/delete、回收站 TrashItem、identify/probe/scrape、download |
auth.ts |
登录、会话探测 SessionInfo、改密、TOTP(TotpStatus/TotpAction)、审计动作枚举 |
browser.ts |
BrowserProvider 合同:会话/标签/快照、快照脱敏类别 SnapshotRedactionKind |
redact.ts |
凭证脱敏的唯一键表与规则:SENSITIVE_KEY_PARTS、SENSITIVE_QUERY_PARAMS、isSensitiveKey、isSensitiveQueryParam、redactString、redactUrl(server logger、Agent 模型上下文、adapters 的 rss/browser 共用) |
api.ts |
Api 端点表(method/path/scope/auth/params/query/body/response;auth:"session" 表示只接受管理员控制台会话)、buildPath()、listEndpoints() |
约定:
- Row schema 与
packages/db/src/types.ts的列一一对应(test/db-sync.test.ts校验);只在 DB 存在的列(密文、key_hash、change_xid)不对外。 - 凭证类输入(
credential、url、plaintext、apiKey)只写不回显。 - schema 不用
.default()做可部分更新对象的缺省(zod 4 在 partial 里仍会填默认值);缺省值放常量。