weichenandCursor
5cbac00b99
Why: Squad runs were one file workspace and one JSON file the leader had to emit, so agents, issues, and retries had no durable queue. Evidence: GOWORK=off go test -count=1 -timeout 180s ./... Impact: /v1/runs stays. New routes mount only when the orchestration database opens. SSE for a run id now skips other runs. External CLIs are not a sandbox. Refs: feature/orchestration-platform S1-S7 Co-authored-by: Cursor <cursoragent@cursor.com>
11 lines
678 B
Go
11 lines
678 B
Go
package platform
|
|
|
|
// 未核实项按保守语义实现，不假装和 multica 的内部表一致：
|
|
//
|
|
// - 权限只有 private 与 public_to 两个档，public_to 用一张允许列表，没有基于角色的工作区 ACL。
|
|
// - 委派深度上限是 4。上游没有写成文的跳数。
|
|
// - Access 的三层存储没有照搬。本库是编排状态；目标引擎仍用原来的文件工作区。
|
|
// - 外部 CLI、扩展和包都以调用者权限运行，不是沙箱。
|
|
// - 看板、Git 托管关联和 cloud runtime 不在这个库里。
|
|
const ConservativeNotes = "private|public_to allow-list; delegation depth 4; file workspace kept for the goal engine"
|