Skip to content

Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.

refactor(session): remove legacy path compatibility (#6642)

Qin Haojie60e48ca4ab0b
diff --git a/bot/vikingbot/openviking_mount/README.md b/bot/vikingbot/openviking_mount/README.md
index 4bc4feb0e..a81eb71f7 100644
--- a/bot/vikingbot/openviking_mount/README.md
+++ b/bot/vikingbot/openviking_mount/README.md
@@ -174,7 +174,7 @@ OpenViking URI 到本地文件路径的映射规则:
OpenViking URI 本地路径
------------------- ------------------
viking://resources/foo -> {mount_point}/resources/foo
-viking://session/bar -> {mount_point}/session/bar
+viking://user/<uid>/sessions/bar -> {mount_point}/user/<uid>/sessions/bar
viking://user/<uid>/baz -> {mount_point}/user/<uid>/baz
```
diff --git a/docs/en/api/05-sessions.md b/docs/en/api/05-sessions.md
index 3fcd6c8f9..b7bee72a6 100644
--- a/docs/en/api/05-sessions.md
+++ b/docs/en/api/05-sessions.md
@@ -12,8 +12,8 @@ viking://user/{user_id}/sessions/{session_id}
```
Session APIs are scoped to the authenticated user and return canonical user
-session URIs. URI-based APIs may also accept the backward-compatible
-`viking://session/{session_id}` alias, resolved in the same user context.
+session URIs. URI-based APIs also accept the `viking://~/sessions/{session_id}`
+home alias for the authenticated user.
## API Reference
diff --git a/docs/en/api/12-content.md b/docs/en/api/12-content.md
index 790558f90..48b668996 100644
--- a/docs/en/api/12-content.md
+++ b/docs/en/api/12-content.md
@@ -150,7 +150,7 @@ Read the complete text of an L0, L1, or L2 file.
- `read()` accepts file URIs only. Passing an existing directory URI returns `INVALID_ARGUMENT` (`400`), not `NOT_FOUND`. This error carries a structured `details` payload — `details.expected` is `"file"`, `details.actual` is `"directory"`, and `details.resource` is the offending URI (present on the HTTP path) — so clients can detect a file-vs-directory mismatch programmatically (for example, fall back to `list`) instead of string-matching the message.
- Instead of a Viking URI, you may pass the 32-character hex `id` returned by `stat()` for a file. The server looks up the URI via the vector index and applies the same permission checks. Because indexing is asynchronous, a newly returned ID might not be resolvable immediately; lookup also fails if the corresponding vector record has been deleted. In both cases, the server returns `NOT_FOUND` and indicates that the data may not have been indexed yet or may have been deleted.
-- Public URI parameters accept `resources`, `user`, and `agent` scopes. For session files, use `viking://user/{user_id}/sessions/{session_id}/messages.jsonl` or the backward-compatible `viking://session/{session_id}/messages.jsonl` alias. Internal scopes such as `temp` and `queue` return `INVALID_URI`.
+- Public URI parameters accept `resources`, `user`, and `agent` scopes. For session files, use `viking://user/{user_id}/sessions/{session_id}/messages.jsonl` or the current user's `viking://~/sessions/{session_id}/messages.jsonl` home alias. The old `viking://session/...` scope and internal scopes such as `temp` and `queue` return `INVALID_URI`.
- Sensitive information handling: default reads (`raw=false`) preserve existing placeholders and sanitize remaining registered plaintext values; `raw=true` skips default sanitization and returns the restored content. Restoration affects only the response, not stored files; see [Privacy Configs](../concepts/13-privacy.md).
**Python SDK**
diff --git a/docs/en/concepts/04-viking-uri.md b/docs/en/concepts/04-viking-uri.md
index eae0fa9c7..6cb67ea8c 100644
--- a/docs/en/concepts/04-viking-uri.md
+++ b/docs/en/concepts/04-viking-uri.md
@@ -24,8 +24,7 @@ viking://{scope}/{path}
| **upload** | Temporary upload files | Temporary | Internal |
Public API and CLI filesystem/content operations accept the public scopes
-`resources`, `user`, and `agent`, plus the root URI `viking://`. `session` is retained
-as a backward-compatible alias for user session paths; new session data lives
+`resources`, `user`, and `agent`, plus the root URI `viking://`. Session data lives
under `viking://user/{user_id}/sessions`.
`temp`, `queue`, and `upload` are internal implementation
scopes and cannot be addressed directly through public API URI parameters.
@@ -170,9 +169,7 @@ viking://user/{user_id}/sessions/{session_id}/history # Archived history
viking://~/sessions/{session_id}/ # Your own session, via the home alias
```
-`viking://session/{session_id}` is accepted as a backward-compatible alias for
-the current user's session path. It is not a separate storage root for new
-session data.
+The old `viking://session/{session_id}` URI is no longer accepted.
## Path Variables
diff --git a/docs/en/concepts/08-session.md b/docs/en/concepts/08-session.md
index 84b3f78a2..163dc95bc 100644
--- a/docs/en/concepts/08-session.md
+++ b/docs/en/concepts/08-session.md
@@ -253,8 +253,7 @@ viking://~/memories/
`viking://user/{user_id}/sessions/{session_id}` for the authenticated caller.
The uid-less spelling `viking://user/sessions/{session_id}` is no longer accepted
and returns an error pointing at the `viking://~/...` form. The old
-`viking://session/{session_id}` form is still accepted as a backward-compatible
-alias for the same session path and is not a separate storage root.
+`viking://session/{session_id}` form is also rejected.
## Related Documents
diff --git a/docs/zh/api/05-sessions.md b/docs/zh/api/05-sessions.md
index 1f1cc4a6a..80ff0149c 100644
--- a/docs/zh/api/05-sessions.md
+++ b/docs/zh/api/05-sessions.md
@@ -12,8 +12,7 @@ viking://user/{user_id}/sessions/{session_id}
```
Session API 按认证用户作用域访问会话,并返回 canonical user session URI。
-基于 URI 的 API 也可以接受向后兼容的 `viking://session/{session_id}` 别名,
-该别名会在同一个用户上下文中解析。
+基于 URI 的 API 可以使用 `viking://~/sessions/{session_id}` 表示当前用户的会话。
## API 参考
diff --git a/docs/zh/api/12-content.md b/docs/zh/api/12-content.md
index 2bc14bdd6..8949e62d1 100644
--- a/docs/zh/api/12-content.md
+++ b/docs/zh/api/12-content.md
@@ -150,7 +150,7 @@ ov overview viking://resources/docs/
- `read()` 只接受文件 URI。传入已存在的目录时返回 `INVALID_ARGUMENT`(HTTP `400`)。HTTP 响应的 `details` 包含 `expected: "file"`、`actual: "directory"` 和目标 `resource` URI,客户端可据此改用目录列表接口,无需解析错误消息。
- 除 Viking URI 外,还可以传入 `stat()` 返回的 32 字符十六进制文件 `id`。服务端通过向量索引查找对应 URI 并执行相同的权限校验。由于索引是异步生成的,新返回的 ID 可能暂时无法解析;对应向量记录被删除后,按 ID 查询也会失败。这两种情况下,服务端都会返回 `NOT_FOUND`,并提示数据可能尚未索引或已经删除。
-- 公开 URI 参数接受 `resources`、`user` 和 `agent` 作用域。访问 session 文件时,使用 `viking://user/{user_id}/sessions/{session_id}/messages.jsonl`,也可以使用向后兼容的 `viking://session/{session_id}/messages.jsonl` 别名。`temp`、`queue` 等内部作用域会返回 `INVALID_URI`。
+- 公开 URI 参数接受 `resources`、`user` 和 `agent` 作用域。访问 session 文件时,使用 `viking://user/{user_id}/sessions/{session_id}/messages.jsonl`,或表示当前用户的 `viking://~/sessions/{session_id}/messages.jsonl`。旧的 `viking://session/...` 以及 `temp`、`queue` 等内部作用域会返回 `INVALID_URI`。
- 敏感信息的处理:默认读取(`raw=false`)保留已有占位符,并对残留的已登记明文脱敏;`raw=true` 跳过默认脱敏,返回加载还原后的内容,该还原只影响返回内容,不修改存储文件,详见[隐私配置](../concepts/13-privacy.md)。
**Python SDK**
diff --git a/docs/zh/concepts/04-viking-uri.md b/docs/zh/concepts/04-viking-uri.md
index 2c8cc6303..6495f60f4 100644
--- a/docs/zh/concepts/04-viking-uri.md
+++ b/docs/zh/concepts/04-viking-uri.md
@@ -24,8 +24,7 @@ viking://{scope}/{path}
| **upload** | 临时上传文件 | 临时 | 内部 |
公开 API 和 CLI 的文件系统/内容操作接受公开作用域 `resources`、`user` 和 `agent`,
-以及根 URI `viking://`。`session` 保留为 user session 路径的向后兼容别名;
-新 session 数据位于 `viking://user/{user_id}/sessions`。
+以及根 URI `viking://`。Session 数据位于 `viking://user/{user_id}/sessions`。
`temp`、`queue` 和 `upload` 是内部实现作用域,不能通过公开 API 的 URI 参数直接访问。
### Home 别名 `~`
@@ -160,8 +159,7 @@ viking://user/{user_id}/sessions/{session_id}/history # 归档历史
viking://~/sessions/{session_id}/ # 自己的会话(家目录别名写法)
```
-`viking://session/{session_id}` 会作为当前用户 session 路径的向后兼容别名被接受。
-它不是新会话数据的独立存储根。
+旧 URI `viking://session/{session_id}` 不再接受。
## 路径变量
diff --git a/docs/zh/concepts/08-session.md b/docs/zh/concepts/08-session.md
index 5949e9674..58e65cbad 100644
--- a/docs/zh/concepts/08-session.md
+++ b/docs/zh/concepts/08-session.md
@@ -252,8 +252,7 @@ viking://~/memories/
`viking://~/sessions/{session_id}` 使用家目录别名,服务端会按认证身份将其展开为
`viking://user/{user_id}/sessions/{session_id}`。无 uid 的写法
`viking://user/sessions/{session_id}` 不再被接受,请求会报错并提示改用 `viking://~/...`。
-`viking://session/{session_id}` 仍会作为同一个 session 路径的向后兼容别名被接受,
-不是独立的存储根。
+旧 URI `viking://session/{session_id}` 也不再接受。
## 相关文档
diff --git a/examples/dsh-memory-plugin/client.test.mjs b/examples/dsh-memory-plugin/client.test.mjs
index ee4a7ef80..27ec3ca3e 100644
--- a/examples/dsh-memory-plugin/client.test.mjs
+++ b/examples/dsh-memory-plugin/client.test.mjs
@@ -14,7 +14,7 @@ test("client sends OpenViking identity headers and preserves response trace ids"
seen = { url, init };
return new Response(JSON.stringify({
status: "ok",
- result: { trace_id: "trace-123", archive_uri: "viking://session/x" },
+ result: { trace_id: "trace-123", archive_uri: "viking://user/default/sessions/x" },
}), {
status: 200,
headers: { "Content-Type": "application/json" },
diff --git a/examples/openclaw-plugin/README_CN.md b/examples/openclaw-plugin/README_CN.md
index 84ed287b3..26414127d 100644
--- a/examples/openclaw-plugin/README_CN.md
+++ b/examples/openclaw-plugin/README_CN.md
@@ -80,7 +80,7 @@ Agent 会自动完成安装 → 配置 → 重启 → 验证。详见 [INSTALL-A
| `ov_list` | 在检索后列出 OpenViking 目录,补查同级切片和 `.overview.md` 文件 |
| `openviking_tool_result_list` | 列出当前 session 中被外置的大工具输出 |
| `openviking_tool_result_search` | 在外置工具输出中按关键词搜索 |
-| `openviking_tool_result_read` | 通过 `viking://session/.../tool-results/...` ref 分页读取外置工具输出 |
+| `openviking_tool_result_read` | 通过 `viking://user/{user_id}/sessions/.../tool-results/...` ref 分页读取外置工具输出 |
插件还提供手动 slash command:`/add-resource`、`/add-skill`、`/ov-search`、`/ov-recall-trace`。Agent 可见的 `add_resource` 工具默认禁用(`enableAddResourceTool=false`),避免搜索/检索阶段误触发资源导入;如确实要允许 Agent 导入资源,可显式设置 `enableAddResourceTool=true`,否则使用手动 `/add-resource`。
@@ -88,7 +88,7 @@ Agent 会自动完成安装 → 配置 → 重启 → 验证。详见 [INSTALL-A
- **发送内容**:每轮 user/assistant 消息文本(已剥离注入的记忆块和元数据噪音)。
- **发送去向**:仅发往你配置的 OpenViking 服务(`baseUrl`)。插件本身只与该服务通信;服务端对 embedding、VLM 等模型的调用取决于服务端配置。
-- **存储位置**:所有数据存储在你的 OpenViking 服务上,命名空间包括 `viking://user/*`、`viking://session/*`、`viking://resources/*` 等。
+- **存储位置**:所有数据存储在你的 OpenViking 服务上,命名空间包括 `viking://user/{user_id}/*`(其中包含 sessions)和 `viking://resources/*` 等。
- **API Key**:通过 `X-API-Key` header 发送,不会被日志记录或转发。
- **多租户隔离**:支持 `accountId`、`userId`。可选的 `peer_role` / `peer_prefix` 控制是否把 OpenClaw 说话人写入 OpenViking `peer_id`,并在数据面使用 `X-OpenViking-Actor-Peer`。
@@ -139,7 +139,7 @@ openclaw config get plugins.slots.contextEngine # 应输出:openviking
- OpenClaw 在左侧,仍然是主运行时;插件并不接管 agent 执行本身。
- 插件中间层把 Hook、Context Engine、Tools、Runtime Manager 四部分合并在一个注册单元里。
- 所有 HTTP 调用最终都走 `OpenVikingClient`,由 client 层统一补 `X-OpenViking-*` 头和路由日志。
-- OpenViking 服务端承接 session、memory、archive 和 Phase 2 抽取,底层存储落在 `viking://user/*`(包含 `viking://user/sessions/*`)、`viking://session/*` 和 `viking://resources/*`。
+- OpenViking 服务端承接 session、memory、archive 和 Phase 2 抽取;session 存储在 `viking://user/{user_id}/sessions/*`。
这套拆分的意义,是让 OpenClaw 继续专注推理与编排,让 OpenViking 成为长期上下文的事实源。
@@ -346,7 +346,7 @@ Resource 导入支持远程 URL、Git URL、本地文件、本地目录和 zip
Recall trace 默认关闭。可通过插件配置 `traceRecall`、`traceRecallPersist`、`traceRecallDir` 等开启,然后使用 `ov_recall_trace` 或 `/ov-recall-trace` 查询。持久化 trace 默认写入 `~/.openclaw/openviking/recall-traces`,并可通过保留天数和查询上限控制扫描范围。
-当 OpenViking 将超大工具输出外置时,可见 preview 会包含 `viking://session/<session_id>/tool-results/<tool_result_id>` 引用。使用 `openviking_tool_result_list` 发现 ref,使用 `openviking_tool_result_search` 定位片段,再使用 `openviking_tool_result_read` 结合 `offset`/`limit` 读取原始内容。
+当 OpenViking 将超大工具输出外置时,可见 preview 会包含 `viking://user/{user_id}/sessions/<session_id>/tool-results/<tool_result_id>` 引用。使用 `openviking_tool_result_list` 发现 ref,使用 `openviking_tool_result_search` 定位片段,再使用 `openviking_tool_result_read` 结合 `offset`/`limit` 读取原始内容。
## 运行模式
diff --git a/examples/openclaw-plugin/docs/openviking-openclaw-plugin-guide.md b/examples/openclaw-plugin/docs/openviking-openclaw-plugin-guide.md
index 7c7e2ab1a..1cbfe0f8d 100644
--- a/examples/openclaw-plugin/docs/openviking-openclaw-plugin-guide.md
+++ b/examples/openclaw-plugin/docs/openviking-openclaw-plugin-guide.md
@@ -209,7 +209,7 @@ transformContext auto recall 流程:
### 5.6 外置 Tool Result 回读
-当 OpenViking 服务端将大工具结果外置为 `viking://session/.../tool-results/...` 时,插件提供:
+当 OpenViking 服务端将大工具结果外置为 `viking://user/{user_id}/sessions/.../tool-results/...` 时,插件提供:
| 工具 | 用途 |
| --- | --- |
@@ -262,7 +262,7 @@ transformContext auto recall 流程:
| --- | --- | --- | --- |
| `POST /api/v1/search/find` | 快速语义检索,不依赖 session context | 自动召回、`memory_recall`、`ov_search`、`memory_forget` | body: `query`、`target_uri`、`limit`、`score_threshold`;返回 `memories[]`、`resources[]`、`skills[]`,每项含 `uri`、`level`、`abstract`、`score`、`category`:`client.ts:428` |
| `POST /api/v1/search/search` | 带 session context 和 intent analysis 的检索 | 暂未使用 | body 可带 `session_id`;返回 `query_plan` / `query_results`。当前插件为了稳定和低延迟统一用 `find()`,session context 由插件自己组装 |
-| `POST /api/v1/search/grep` | 正则/关键词内容搜索 | `ov_archive_search` | body: `uri`、`pattern`、`case_insensitive`、`node_limit`;插件限定在 `viking://session/{id}/history` 内搜 archive:`client.ts:897` |
+| `POST /api/v1/search/grep` | 正则/关键词内容搜索 | `ov_archive_search` | body: `uri`、`pattern`、`case_insensitive`、`node_limit`;插件限定在 `viking://user/{user_id}/sessions/{id}/history` 内搜 archive:`client.ts:897` |
| `POST /api/v1/search/glob` | glob 文件匹配 | 暂未封装 | body: `pattern`、`uri`、`node_limit`;适合按 `**/*.md`、`src/**/*.ts` 找资源路径 |
#### 6.2.3 Filesystem / Content
@@ -339,8 +339,8 @@ transformContext auto recall 流程:
| `viking://user/memories` | 当前用户长期记忆别名 |
| `viking://resources` | account/resource 知识库 |
| `viking://user/skills` | 当前 agent skill 空间 |
-| `viking://session/{sessionId}/history` | session archive 历史 |
-| `viking://session/{sessionId}/tool-results/{id}` | 外置工具结果 |
+| `viking://user/{user_id}/sessions/{sessionId}/history` | session archive 历史 |
+| `viking://user/{user_id}/sessions/{sessionId}/tool-results/{id}` | 外置工具结果 |
插件通过 `viking://user/...` 写入和检索 user-scoped memory;OpenViking 会根据请求里的租户身份和 actor peer context 解析这个别名。agent 维度通过 `peer_id` / `X-OpenViking-Actor-Peer` 表达,不再使用旧 agent URI namespace。
@@ -1322,7 +1322,7 @@ python health_check_tools/ov-healthcheck.py
5. 新开一轮问“我之前偏好什么语言回复技术文档?”。
6. 观察最新 user message 是否注入 `<relevant-memories>`,或用 `memory_recall` 显式查。
7. 用 OpenViking Console/TUI 检查 `viking://user/.../memories` 是否产生 leaf memory。
-8. 对长工具输出场景,确认 preview 中有 `viking://session/.../tool-results/...`,再用 tool-result 工具读取完整内容。
+8. 对长工具输出场景,确认 preview 中有 `viking://user/{user_id}/sessions/.../tool-results/...`,再用 tool-result 工具读取完整内容。
---
diff --git a/examples/openclaw-plugin/docs/openviking-websocket-rpc-api.md b/examples/openclaw-plugin/docs/openviking-websocket-rpc-api.md
index ce10d7896..ebf3510f0 100644
--- a/examples/openclaw-plugin/docs/openviking-websocket-rpc-api.md
+++ b/examples/openclaw-plugin/docs/openviking-websocket-rpc-api.md
@@ -450,7 +450,7 @@ Inspect recall traces when `traceRecall` is enabled.
"name": "openviking_tool_result_search",
"sessionKey": "main",
"args": {
- "tool_output_ref": "viking://session/<session_id>/tool-results/<tool_result_id>",
+ "tool_output_ref": "viking://user/{user_id}/sessions/<session_id>/tool-results/<tool_result_id>",
"query": "error",
"limit": 10,
"context_chars": 300
@@ -463,7 +463,7 @@ Inspect recall traces when `traceRecall` is enabled.
| 字段 | 类型 | 必填 | 说明 |
|------|------|------|------|
-| `tool_output_ref` | string | 是 | `viking://session/.../tool-results/...` URI,也兼容 `ref` 或 `uri` |
+| `tool_output_ref` | string | 是 | `viking://user/{user_id}/sessions/.../tool-results/...` URI,也兼容 `ref` 或 `uri` |
| `query` | string | 是 | 搜索关键词或精确文本 |
| `limit` | number | 否 | 最多匹配数,默认 20 |
| `context_chars` | number | 否 | 每个命中周围保留字符数,默认 300,也兼容 `contextChars` |
@@ -481,7 +481,7 @@ Inspect recall traces when `traceRecall` is enabled.
"name": "openviking_tool_result_read",
"sessionKey": "main",
"args": {
- "tool_output_ref": "viking://session/<session_id>/tool-results/<tool_result_id>",
+ "tool_output_ref": "viking://user/{user_id}/sessions/<session_id>/tool-results/<tool_result_id>",
"offset": 0,
"limit": 4000
}
@@ -493,7 +493,7 @@ Inspect recall traces when `traceRecall` is enabled.
| 字段 | 类型 | 必填 | 说明 |
|------|------|------|------|
-| `tool_output_ref` | string | 是 | `viking://session/.../tool-results/...` URI,也兼容 `ref` 或 `uri` |
+| `tool_output_ref` | string | 是 | `viking://user/{user_id}/sessions/.../tool-results/...` URI,也兼容 `ref` 或 `uri` |
| `offset` | number | 否 | 起始字符偏移,默认 0 |
| `limit` | number | 否 | 最多返回字符数 |
diff --git a/examples/openclaw-plugin/memory-ranking.ts b/examples/openclaw-plugin/memory-ranking.ts
index 38cbad323..be7aaa066 100644
--- a/examples/openclaw-plugin/memory-ranking.ts
+++ b/examples/openclaw-plugin/memory-ranking.ts
@@ -231,7 +231,7 @@ function lexicalOverlapBoost(tokens: string[], text: string, maxBoost = DEFAULT_
function inferResourceType(item: FindResultItem): string | undefined {
if (item.uri.startsWith("viking://resources")) return "resource";
- if (item.uri.startsWith("viking://session/")) return "session";
+ if (/^viking:\/\/(?:user\/[^/]+|~)\/sessions(?:\/|$)/.test(item.uri)) return "session";
if (item.uri.startsWith("viking://user/skills") || item.uri.startsWith("viking://~/skills") || item.uri.startsWith("viking://skills")) return "agent";
if (item.uri.startsWith("viking://user/") || item.uri.startsWith("viking://~/")) return "user";
return undefined;
diff --git a/examples/openclaw-plugin/plugin/openviking-archive-tools.ts b/examples/openclaw-plugin/plugin/openviking-archive-tools.ts
index 6bb11a21a..6b336bf2c 100644
--- a/examples/openclaw-plugin/plugin/openviking-archive-tools.ts
+++ b/examples/openclaw-plugin/plugin/openviking-archive-tools.ts
@@ -150,7 +150,7 @@ export function registerOpenVikingArchiveTools(deps: OpenVikingArchiveToolsDeps)
trigger: { query, derivedKeywords: [query] },
searches: [{
resourceType: "archive",
- targetUriResolved: archiveId ? `viking://session/${ovSessionId}/history/${archiveId}` : `viking://session/${ovSessionId}/history`,
+ targetUriResolved: archiveId ? `viking://~/sessions/${ovSessionId}/history/${archiveId}` : `viking://~/sessions/${ovSessionId}/history`,
limit: deps.traceRecallMaxResultsPerSearch,
durationMs: Date.now() - started,
total: result.matches?.length ?? result.count ?? 0,
diff --git a/examples/openclaw-plugin/plugin/openviking-recall-trace-runtime.ts b/examples/openclaw-plugin/plugin/openviking-recall-trace-runtime.ts
index e24468ce0..8fe3090e7 100644
--- a/examples/openclaw-plugin/plugin/openviking-recall-trace-runtime.ts
+++ b/examples/openclaw-plugin/plugin/openviking-recall-trace-runtime.ts
@@ -119,7 +119,7 @@ function toLimitedInteger(value: unknown, fallback: number, min: number, max: nu
function inferUriType(uri: string): string {
if (uri.startsWith("viking://resources") || uri.startsWith("viking://resource")) return "resource";
- if (uri.startsWith("viking://session/")) return "session";
+ if (/^viking:\/\/(?:user\/[^/]+|~)\/sessions(?:\/|$)/.test(uri)) return "session";
if (uri.startsWith("viking://user/skills") || uri.startsWith("viking://~/skills") || uri.startsWith("viking://skills")) return "skill";
if (uri.startsWith("viking://user/") || uri.startsWith("viking://~/")) return "user_memory";
if (uri.includes("/archive") || uri.includes("/history/")) return "archive";
diff --git a/examples/openclaw-plugin/plugin/openviking-tool-result-tools.ts b/examples/openclaw-plugin/plugin/openviking-tool-result-tools.ts
index d1896d299..2227ec7a0 100644
--- a/examples/openclaw-plugin/plugin/openviking-tool-result-tools.ts
+++ b/examples/openclaw-plugin/plugin/openviking-tool-result-tools.ts
@@ -78,7 +78,7 @@ export function parseToolResultRef(value: unknown): ToolResultRef | null {
if (!raw) {
return null;
}
- const match = raw.match(/^viking:\/\/session\/([^/]+)\/tool-results\/([^/?#]+)(?:[?#].*)?$/);
+ const match = raw.match(/^viking:\/\/(?:user\/[^/]+|~)\/sessions\/([^/]+)\/tool-results\/([^/?#]+)(?:[?#].*)?$/);
if (!match) {
return null;
}
@@ -90,7 +90,7 @@ export function parseToolResultRef(value: unknown): ToolResultRef | null {
return {
sessionId,
toolResultId,
- ref: `viking://session/${encodeURIComponent(sessionId)}/tool-results/${encodeURIComponent(toolResultId)}`,
+ ref: raw.split(/[?#]/, 1)[0]!,
};
}
@@ -120,14 +120,14 @@ export function registerOpenVikingToolResultTools({
description:
"Restore the full original content of a tool result that was externalized by OpenViking. " +
"Use when a previous tool result was externalized and only a preview is visible — " +
- "the preview contains a [tool-result-ref] or viking://session/.../tool-results/... URI. " +
+ "the preview contains a [tool-result-ref] or viking://user/.../sessions/.../tool-results/... URI. " +
"\"Read\" tool returns the same truncated preview; this tool returns the complete content. " +
"To read all content: pass offset=0 and a limit large enough to cover the whole result " +
"(e.g. limit=100000). Use offset/limit for paging only when you need a specific section.",
parameters: Type.Object({
tool_output_ref: Type.String({
description:
- "Exact OV URI from the preview, e.g. viking://session/<session_id>/tool-results/<tool_result_id>",
+ "Exact OV URI from the preview, e.g. viking://user/<user_id>/sessions/<session_id>/tool-results/<tool_result_id>",
}),
offset: Type.Optional(Type.Number({ description: "Unicode character offset. Default: 0" })),
limit: Type.Optional(Type.Number({ description: "Maximum Unicode characters to read. Default: 20000" })),
@@ -147,7 +147,7 @@ export function registerOpenVikingToolResultTools({
const parsed = parseToolResultRef(params.tool_output_ref ?? params.ref ?? params.uri);
if (!parsed) {
return {
- content: [{ type: "text", text: "Error: tool_output_ref must be a viking://session/.../tool-results/... URI." }],
+ content: [{ type: "text", text: "Error: tool_output_ref must be a viking://user/.../sessions/.../tool-results/... URI." }],
details: { error: "invalid_tool_output_ref" },
};
}
@@ -221,7 +221,7 @@ export function registerOpenVikingToolResultTools({
parameters: Type.Object({
tool_output_ref: Type.String({
description:
- "Exact OV URI from the preview, e.g. viking://session/<session_id>/tool-results/<tool_result_id>",
+ "Exact OV URI from the preview, e.g. viking://user/<user_id>/sessions/<session_id>/tool-results/<tool_result_id>",
}),
query: Type.String({ description: "Keyword or exact text to search for" }),
limit: Type.Optional(Type.Number({ description: "Maximum matches. Default: 20" })),
@@ -242,7 +242,7 @@ export function registerOpenVikingToolResultTools({
const parsed = parseToolResultRef(params.tool_output_ref ?? params.ref ?? params.uri);
if (!parsed) {
return {
- content: [{ type: "text", text: "Error: tool_output_ref must be a viking://session/.../tool-results/... URI." }],
+ content: [{ type: "text", text: "Error: tool_output_ref must be a viking://user/.../sessions/.../tool-results/... URI." }],
details: { error: "invalid_tool_output_ref" },
};
}
diff --git a/examples/openclaw-plugin/skills/install-openviking-memory/SKILL.md b/examples/openclaw-plugin/skills/install-openviking-memory/SKILL.md
index 162f99dce..21ab175ee 100644
--- a/examples/openclaw-plugin/skills/install-openviking-memory/SKILL.md
+++ b/examples/openclaw-plugin/skills/install-openviking-memory/SKILL.md
@@ -587,7 +587,7 @@ Requires `traceRecall=true`; persisted lookup also requires `traceRecallPersist=
### Externalized Tool Result Tools
-Use these when a preview contains a `viking://session/<session_id>/tool-results/<tool_result_id>` ref.
+Use these when a preview contains a `viking://user/{user_id}/sessions/<session_id>/tool-results/<tool_result_id>` ref.
| Tool | Parameters |
|---|---|
diff --git a/examples/openclaw-plugin/skills/openviking-context-database/SKILL.md b/examples/openclaw-plugin/skills/openviking-context-database/SKILL.md
index 18b9e1265..61ed7f498 100644
--- a/examples/openclaw-plugin/skills/openviking-context-database/SKILL.md
+++ b/examples/openclaw-plugin/skills/openviking-context-database/SKILL.md
@@ -242,7 +242,7 @@ Trace records exist only when `traceRecall=true`; persisted lookup requires `tra
### Externalized tool-result tools
-Use when a preview contains `viking://session/<session_id>/tool-results/<tool_result_id>`.
+Use when a preview contains `viking://user/{user_id}/sessions/<session_id>/tool-results/<tool_result_id>`.
| Tool | Parameters |
|---|---|
diff --git a/examples/openclaw-plugin/tests/ut/plugin-modules.test.ts b/examples/openclaw-plugin/tests/ut/plugin-modules.test.ts
index 64ad4927b..f5e970944 100644
--- a/examples/openclaw-plugin/tests/ut/plugin-modules.test.ts
+++ b/examples/openclaw-plugin/tests/ut/plugin-modules.test.ts
@@ -581,7 +581,7 @@ describe("plugin module seams", () => {
const readFactory = registerTool.mock.calls[0]?.[0];
const readTool = readFactory({ sessionId: "session-1" });
const result = await readTool.execute("call-1", {
- tool_output_ref: "viking://session/ov-session-1/tool-results/tr_1",
+ tool_output_ref: "viking://user/default/sessions/ov-session-1/tool-results/tr_1",
});
expect(readToolResult).toHaveBeenCalledWith(
@@ -599,7 +599,7 @@ describe("plugin module seams", () => {
const grepSessionArchives = vi.fn().mockResolvedValue({
count: 1,
matches: [{
- uri: "viking://session/ov-session-1/history/archive_001#L12",
+ uri: "viking://user/default/sessions/ov-session-1/history/archive_001#L12",
line: 12,
content: "important command output",
}],
diff --git a/examples/openclaw-plugin/tests/ut/tool-round-trip.test.ts b/examples/openclaw-plugin/tests/ut/tool-round-trip.test.ts
index f9dc9170e..8cf6b48d3 100644
--- a/examples/openclaw-plugin/tests/ut/tool-round-trip.test.ts
+++ b/examples/openclaw-plugin/tests/ut/tool-round-trip.test.ts
@@ -145,7 +145,7 @@ describe("convertToAgentMessages: structured tool round-trip", () => {
tool_status: "completed",
tool_input: { path: "/tmp/big.txt" },
tool_output: "preview only",
- tool_output_ref: "viking://session/s1/tool-results/tr_call_big_abc",
+ tool_output_ref: "viking://user/default/sessions/s1/tool-results/tr_call_big_abc",
tool_output_original_chars: 120000,
},
],
@@ -155,7 +155,7 @@ describe("convertToAgentMessages: structured tool round-trip", () => {
const toolResult = result[1] as Record<string, unknown>;
const content = toolResult.content as Array<Record<string, string>>;
expect(content[0]!.text).toContain("preview only");
- expect(content[0]!.text).toContain("viking://session/s1/tool-results/tr_call_big_abc");
+ expect(content[0]!.text).toContain("viking://user/default/sessions/s1/tool-results/tr_call_big_abc");
expect(content[0]!.text).toContain("original_chars=120000");
});
diff --git a/examples/openclaw-plugin/tests/ut/tools.test.ts b/examples/openclaw-plugin/tests/ut/tools.test.ts
index 6d0fdfbda..a04026e26 100644
--- a/examples/openclaw-plugin/tests/ut/tools.test.ts
+++ b/examples/openclaw-plugin/tests/ut/tools.test.ts
@@ -711,7 +711,7 @@ describe("Tool: OpenViking tool result access", () => {
total_chars: 42,
has_more: true,
metadata: {
- storage_uri: "viking://session/test-session/tool-results/tr_call_abc",
+ storage_uri: "viking://user/default/sessions/test-session/tool-results/tr_call_abc",
tool_name: "read_file",
},
});
@@ -723,7 +723,7 @@ describe("Tool: OpenViking tool result access", () => {
const tool = tools.get("openviking_tool_result_read")!;
const result = await tool.execute("tc-read", {
- tool_output_ref: "viking://session/test-session/tool-results/tr_call_abc",
+ tool_output_ref: "viking://user/default/sessions/test-session/tool-results/tr_call_abc",
offset: 5,
limit: 10,
}) as ToolResult;
@@ -731,7 +731,7 @@ describe("Tool: OpenViking tool result access", () => {
expect(result.content[0]!.text).toBe("raw");
expect(result.details).toMatchObject({
action: "read",
- tool_output_ref: "viking://session/test-session/tool-results/tr_call_abc",
+ tool_output_ref: "viking://user/default/sessions/test-session/tool-results/tr_call_abc",
tool_result_id: "tr_call_abc",
offset: 5,
limit: 10,
@@ -766,7 +766,7 @@ describe("Tool: OpenViking tool result access", () => {
const tool = tools.get("openviking_tool_result_search")!;
const result = await tool.execute("tc-search", {
- tool_output_ref: "viking://session/test-session/tool-results/tr_call_abc",
+ tool_output_ref: "viking://user/default/sessions/test-session/tool-results/tr_call_abc",
query: "needle",
limit: 2,
context_chars: 15,
@@ -777,7 +777,7 @@ describe("Tool: OpenViking tool result access", () => {
expect(result.content[0]!.text).toContain("hay needle stack");
expect(result.details).toMatchObject({
action: "searched",
- tool_output_ref: "viking://session/test-session/tool-results/tr_call_abc",
+ tool_output_ref: "viking://user/default/sessions/test-session/tool-results/tr_call_abc",
tool_result_id: "tr_call_abc",
query: "needle",
match_count: 1,
@@ -793,7 +793,7 @@ describe("Tool: OpenViking tool result access", () => {
tool_results: [
{
tool_result_id: "tr_call_abc",
- storage_uri: "viking://session/test-session/tool-results/tr_call_abc",
+ storage_uri: "viking://user/default/sessions/test-session/tool-results/tr_call_abc",
tool_name: "read_file",
original_chars: 42000,
created_at: "2026-05-15T00:00:00Z",
@@ -814,7 +814,7 @@ describe("Tool: OpenViking tool result access", () => {
expect(result.content[0]!.text).toContain("read_file");
expect(result.content[0]!.text).toContain("original_chars=42000");
- expect(result.content[0]!.text).toContain("viking://session/test-session/tool-results/tr_call_abc");
+ expect(result.content[0]!.text).toContain("viking://user/default/sessions/test-session/tool-results/tr_call_abc");
expect(result.details).toMatchObject({
action: "listed",
session_id: "test-session",
@@ -829,7 +829,7 @@ describe("Tool: OpenViking tool result access", () => {
const tool = tools.get("openviking_tool_result_read")!;
const result = await tool.execute("tc-read", {
- tool_output_ref: "viking://session/other-session/tool-results/tr_call_abc",
+ tool_output_ref: "viking://user/default/sessions/other-session/tool-results/tr_call_abc",
}) as ToolResult;
expect(result.content[0]!.text).toContain("another session");
@@ -2021,7 +2021,7 @@ describe("Tool: ov_recall_trace", () => {
return okResponse({
matches: [{
line: 12,
- uri: "viking://session/test-session/history/archive_001#L12",
+ uri: "viking://user/default/sessions/test-session/history/archive_001#L12",
content: "discussion about recall traces",
}],
count: 1,
@@ -2045,7 +2045,7 @@ describe("Tool: ov_recall_trace", () => {
const entry = (result.details.entries as any[])[0];
expect(entry.operationType).toBe("archive_grep");
expect(entry.selected).toEqual(expect.arrayContaining([
- expect.objectContaining({ uri: "viking://session/test-session/history/archive_001#L12", displayed: true }),
+ expect.objectContaining({ uri: "viking://user/default/sessions/test-session/history/archive_001#L12", displayed: true }),
]));
});
});
diff --git a/examples/pi-experimental-context-management/CONTEXT-WINDOW.md b/examples/pi-experimental-context-management/CONTEXT-WINDOW.md
index da91a555e..c5b0c5137 100644
--- a/examples/pi-experimental-context-management/CONTEXT-WINDOW.md
+++ b/examples/pi-experimental-context-management/CONTEXT-WINDOW.md
@@ -626,9 +626,9 @@ summarizer.
`<archive_uri>/.overview.md` and, as a secondary signal, `GET /tasks/{id}`.
`GET /sessions/{id}/context` is not used for waiting — right after a commit it
still returns the previous overview.
-- **No `viking://~` alias.** The server rejects it with `Invalid scope '~'`.
- URIs come from `GET /sessions/{id}`'s `uri` field (cached) or from the
- commit's `archive_uri`; the fallback is the legacy `viking://session/{id}`.
+- Session URIs come from `GET /sessions/{id}`'s `uri` field (cached) or from the
+ commit's `archive_uri`; the fallback is the current user's
+ `viking://~/sessions/{id}` home alias.
- **`ctx.getContextUsage()` is stale for one turn after a reset.** It counts
the untransformed session, so the core ignores it while
`awaitingFirstObservation` is set and reports its own estimate as
diff --git a/examples/pi-experimental-context-management/client.ts b/examples/pi-experimental-context-management/client.ts
index 764d8621c..f3dbb4eae 100644
--- a/examples/pi-experimental-context-management/client.ts
+++ b/examples/pi-experimental-context-management/client.ts
@@ -323,11 +323,8 @@ export class OVClient {
* Session root URI, e.g. `viking://user/default/sessions/pi-abc`.
*
* Reads the canonical `uri` from `GET /api/v1/sessions/{id}` and caches it.
- * When the server does not answer (or answers without a `viking://` uri) this
- * falls back to the legacy alias `viking://session/{id}`, which the target
- * server still accepts. The fallback is deliberately NOT cached so a later
- * call can still pick up the canonical URI. Note the `viking://~` alias is
- * rejected by the server ("Invalid scope '~'") and is never produced here.
+ * If the metadata request fails, use the current user's home alias. The
+ * fallback is not cached so a later call can still pick up the canonical URI.
*/
async sessionRootUri(sessionId: string): Promise<string> {
const id = String(sessionId ?? "");
@@ -341,9 +338,9 @@ export class OVClient {
return uri;
}
} catch {
- // fall through to the legacy alias
+ // fall through to the home alias
}
- return `viking://session/${id}`;
+ return `viking://~/sessions/${id}`;
}
/** Strip `/history/archive_NNN` off a commit's `archive_uri` to get the session root. */
diff --git a/examples/pi-experimental-context-management/scripts/e2e-window.mjs b/examples/pi-experimental-context-management/scripts/e2e-window.mjs
index 350737047..989866633 100755
--- a/examples/pi-experimental-context-management/scripts/e2e-window.mjs
+++ b/examples/pi-experimental-context-management/scripts/e2e-window.mjs
@@ -420,7 +420,7 @@ async function sessionRootUri(ovSessionId) {
const res = await ovFetch(`/api/v1/sessions/${encodeURIComponent(ovSessionId)}`);
const uri = typeof res.body?.result?.uri === "string" ? res.body.result.uri.trim() : "";
if (uri.startsWith("viking://")) return uri.replace(/\/+$/, "");
- return `viking://session/${ovSessionId}`;
+ return `viking://~/sessions/${ovSessionId}`;
}
/** Archive directories under `<root>/history`, newest first. */
diff --git a/examples/pi-experimental-context-management/tools.ts b/examples/pi-experimental-context-management/tools.ts
index c0a17e014..a2ef7987c 100644
--- a/examples/pi-experimental-context-management/tools.ts
+++ b/examples/pi-experimental-context-management/tools.ts
@@ -26,7 +26,7 @@ export function registerTools(pi: any, client: OVClient, sync?: SyncManager): vo
// so the examples name the two forms it does accept.
description:
"Viking URI prefix to scope search, e.g. 'viking://user/<space>/memories/' " +
- "or 'viking://session/<session id>/'",
+ "or 'viking://~/sessions/<session id>/'",
})),
limit: Type.Optional(Type.Number({ description: "Max results (default: 10)" })),
}),
diff --git a/openviking/core/namespace.py b/openviking/core/namespace.py
index 7cdef5329..e469f0a13 100644
--- a/openviking/core/namespace.py
+++ b/openviking/core/namespace.py
@@ -176,8 +176,6 @@ def canonical_session_uri(ctx: RequestContext, session_id: Optional[str] = None)
def is_session_uri(uri: str) -> bool:
parts = uri_parts(uri)
- if parts[:1] == ["session"]:
- return True
return len(parts) >= 3 and parts[0] == "user" and parts[2] == "sessions"
@@ -258,14 +256,14 @@ def resolve_uri(
def resolve_request_uri(uri: str, ctx: RequestContext) -> str:
"""Resolve supported URI aliases at an authenticated request boundary.
- Supported aliases are the ``~`` home alias and the legacy ``session`` scope.
+ The ``~`` home alias resolves to the authenticated user's namespace.
The uid-less ``viking://user/<reserved>`` shorthand is no longer expanded:
it fails closed with a hint pointing at ``viking://~/...``.
"""
# Every authenticated request context carries an effective user identity,
# including ROOT contexts produced by dev, API-key, and trusted auth modes.
- # Resolve only the unambiguous home alias for every role; preserve the
- # existing role-dependent handling of legacy/ambiguous spellings below.
+ # Resolve the home alias for every role; keep the role-dependent handling
+ # of ambiguous uid-less user paths below.
parts = uri_parts(uri)
if parts and parts[0] == "~":
return resolve_current_user_uri(uri, ctx)
@@ -277,8 +275,7 @@ def resolve_request_uri(uri: str, ctx: RequestContext) -> str:
def resolve_current_user_uri(uri: str, ctx: RequestContext) -> str:
"""Resolve a URI field whose contract explicitly denotes the current user.
- Supported aliases are ``viking://~`` (the home alias) and the legacy
- ``viking://session/...`` scope. ``viking://user/<reserved-segment>`` used to
+ The supported alias is ``viking://~``. ``viking://user/<reserved-segment>`` used to
expand to the caller's space; it now fails closed with a corrective hint,
because the same spelling is a valid explicit-uid URI for a user literally
named after the reserved segment.
@@ -292,14 +289,6 @@ def resolve_current_user_uri(uri: str, ctx: RequestContext) -> str:
return canonical_user_root(ctx)
return f"{canonical_user_root(ctx)}/{'/'.join(parts[1:])}"
- if parts[0] == "session":
- if len(parts) == 1:
- return canonical_session_root(ctx)
- canonical = canonical_session_uri(ctx, parts[1])
- if len(parts) > 2:
- canonical = f"{canonical}/{'/'.join(parts[2:])}"
- return canonical
-
if (
parts[0] == "user"
and len(parts) >= 2
diff --git a/openviking/core/uri_validation.py b/openviking/core/uri_validation.py
index 5267fa259..c86fb2b17 100644
--- a/openviking/core/uri_validation.py
+++ b/openviking/core/uri_validation.py
@@ -15,9 +15,11 @@ from openviking_cli.exceptions import InvalidURIError, PermissionDeniedError
from openviking_cli.utils.uri import VikingURI
_PUBLIC_API_SCOPES = frozenset(
- {"", *VikingURI.PUBLIC_SCOPES, *VikingURI.LEGACY_SCOPES, *VikingURI.ALIAS_SCOPES}
+ {"", *VikingURI.PUBLIC_SCOPES, *VikingURI.ALIAS_SCOPES}
+)
+_ALL_API_SCOPES = frozenset(
+ {"", *VikingURI.PUBLIC_SCOPES, *VikingURI.INTERNAL_SCOPES, *VikingURI.ALIAS_SCOPES}
)
-_ALL_API_SCOPES = frozenset({"", *VikingURI.VISITABLE_SCOPES})
def _allowed_api_scopes(
diff --git a/openviking/storage/viking_fs/_access.py b/openviking/storage/viking_fs/_access.py
index 41163a0e3..d0a4f5a78 100644
--- a/openviking/storage/viking_fs/_access.py
+++ b/openviking/storage/viking_fs/_access.py
@@ -4,7 +4,6 @@
import hashlib
import inspect
-import json
import re
from contextlib import contextmanager
from typing import Any, Callable, Dict, List, Mapping, Optional, Sequence, Union
@@ -587,17 +586,8 @@ class _AccessMixin:
):
"""Yield one visible tree page after namespace and ACL filtering."""
real_ctx = self._ctx_or_default(ctx)
- primary_path = self._uri_to_path(uri, ctx=ctx)
- path: Optional[str] = None
- for candidate_path in self._read_paths(uri, ctx=ctx):
- if not await self._read_path_visible(uri, candidate_path, primary_path, real_ctx):
- continue
- if await self._agfs_path_exists(candidate_path):
- path = candidate_path
- break
- if path is None:
- if self._is_session_root_uri(uri):
- return
+ path = self._uri_to_path(uri, ctx=ctx)
+ if not await self._agfs_path_exists(path):
raise NotFoundError(uri, "directory")
if node_limit == 0:
@@ -633,8 +623,6 @@ class _AccessMixin:
acl_enabled=acl_enabled,
):
continue
- if not await self._read_path_visible(uri, entry["path"], primary_path, real_ctx):
- continue
entry_uri = self._alias_uri_for_path(
request_uri=uri,
base_path=path,
@@ -731,64 +719,6 @@ class _AccessMixin:
safe_parts = [self._shorten_component(p, self._MAX_FILENAME_BYTES) for p in parts]
return f"/local/{account_id}/{'/'.join(safe_parts)}"
- def _legacy_session_path(self, uri: str, ctx: Optional[RequestContext] = None) -> str:
- """Map a legacy viking://session URI to its pre-user-namespace path."""
- real_ctx = self._ctx_or_default(ctx)
- parts = self._safe_uri_parts(uri)
- safe_parts = [self._shorten_component(p, self._MAX_FILENAME_BYTES) for p in parts]
- return f"/local/{real_ctx.account_id}/{'/'.join(safe_parts)}"
-
- def _legacy_user_session_path(
- self, uri: str, ctx: Optional[RequestContext] = None
- ) -> Optional[str]:
- """Return the legacy nested /session/{user_id}/{session_id} candidate."""
- real_ctx = self._ctx_or_default(ctx)
- parts = self._safe_uri_parts(uri)
- if len(parts) <= 3 or parts[0] != "user" or parts[2] != "sessions":
- return None
- nested_parts = ["session", parts[1], *parts[3:]]
- safe_parts = [self._shorten_component(p, self._MAX_FILENAME_BYTES) for p in nested_parts]
- return f"/local/{real_ctx.account_id}/{'/'.join(safe_parts)}"
-
- def _legacy_session_alias(self, uri: str) -> Optional[str]:
- """Return the old storage alias for a canonical user session URI."""
- parts = self._safe_uri_parts(uri)
- if len(parts) < 3 or parts[0] != "user" or parts[2] != "sessions":
- return None
- suffix = parts[3:]
- return "viking://session" + (f"/{'/'.join(suffix)}" if suffix else "")
-
- def _is_session_root_uri(self, uri: str) -> bool:
- return self._legacy_session_alias(uri) == "viking://session"
-
- def _read_paths(self, uri: str, ctx: Optional[RequestContext] = None) -> List[str]:
- """Return read candidates for a URI, including legacy alias fallbacks."""
- paths = [self._uri_to_path(uri, ctx=ctx)]
-
- legacy_uri = self._legacy_session_alias(uri)
- if legacy_uri:
- for candidate in (
- self._legacy_session_path(legacy_uri, ctx=ctx),
- self._legacy_user_session_path(uri, ctx=ctx),
- ):
- if candidate and candidate not in paths:
- paths.append(candidate)
- return paths
-
- async def _read_path_visible(
- self,
- request_uri: str,
- path: str,
- primary_path: str,
- ctx: RequestContext,
- ) -> bool:
- if path == primary_path:
- return True
- if self._legacy_session_alias(request_uri):
- owner_user_id = self._safe_uri_parts(request_uri)[1]
- return await self._legacy_session_path_visible(path, owner_user_id=owner_user_id)
- return True
-
def _alias_uri_for_path(
self,
*,
@@ -804,8 +734,6 @@ class _AccessMixin:
if entry_path.startswith(base):
separator = "" if request_root.endswith("://") else "/"
candidate_uri = request_root if not rel_path else f"{request_root}{separator}{rel_path}"
- if self._legacy_session_alias(request_uri):
- return candidate_uri
if preserve_request_alias:
return candidate_uri
try:
@@ -824,152 +752,6 @@ class _AccessMixin:
return False
raise
- async def _looks_like_legacy_session_dir(self, path: str) -> bool:
- for leaf in (".meta.json", "messages.jsonl", "history", "tool-results", "tools"):
- if await self._agfs_path_exists(f"{path}/{leaf}"):
- return True
- return False
-
- async def _legacy_session_owner(self, session_root_path: str) -> str:
- try:
- raw = self._handle_agfs_read(
- await self._async_agfs.read(f"{session_root_path}/.meta.json")
- )
- except Exception as exc:
- if is_not_found_error(exc):
- return ""
- raise
- try:
- data = json.loads(self._decode_bytes(raw))
- except json.JSONDecodeError:
- return ""
- if not isinstance(data, dict):
- return ""
- for key in ("created_by_user_id", "user_id", "owner_user_id", "created_by"):
- value = data.get(key)
- if isinstance(value, str) and value.strip():
- return value.strip()
- return ""
-
- async def _legacy_session_visible(
- self,
- session_root_path: str,
- ctx: RequestContext,
- *,
- owner_hint: Optional[str] = None,
- ) -> bool:
- if ctx.role == Role.ROOT:
- return True
- owner = await self._legacy_session_owner(session_root_path)
- if owner:
- return owner == ctx.user.user_id
- if owner_hint:
- return owner_hint == ctx.user.user_id
- return True
-
- async def _legacy_session_path_visible(
- self,
- path: str,
- *,
- owner_user_id: str,
- ) -> bool:
- parts = [p for p in path.strip("/").split("/") if p]
- try:
- session_index = parts.index("session")
- except ValueError:
- return True
- suffix = parts[session_index + 1 :]
- if not suffix:
- return True
-
- root_prefix = "/" + "/".join(parts[: session_index + 1])
- direct_root = f"{root_prefix}/{suffix[0]}"
- if await self._looks_like_legacy_session_dir(direct_root):
- owner = await self._legacy_session_owner(direct_root)
- return bool(owner) and owner == owner_user_id
-
- if len(suffix) >= 2:
- if suffix[0] != owner_user_id:
- return False
- nested_root = f"{root_prefix}/{suffix[0]}/{suffix[1]}"
- if await self._looks_like_legacy_session_dir(nested_root):
- owner = await self._legacy_session_owner(nested_root)
- return not owner or owner == owner_user_id
- return True
-
- async def _legacy_session_root_items(
- self,
- path: str,
- ctx: RequestContext,
- output_root_uri: str,
- owner_user_id: str,
- ) -> List[tuple[Dict[str, Any], str]]:
- try:
- entries = await self._ls_entries(path)
- except Exception as exc:
- if is_not_found_error(exc):
- return []
- raise
-
- items: List[tuple[Dict[str, Any], str]] = []
- for entry in entries:
- name = entry.get("name", "")
- if not name or name in {".", ".."} or not entry.get("isDir"):
- continue
- child_path = f"{path.rstrip('/')}/{name}"
- if await self._looks_like_legacy_session_dir(child_path):
- legacy_owner = await self._legacy_session_owner(child_path)
- if legacy_owner == owner_user_id:
- items.append((entry, f"{output_root_uri}/{name}"))
- continue
-
- if name != owner_user_id:
- continue
- try:
- nested_entries = await self._ls_entries(child_path)
- except Exception as exc:
- if is_not_found_error(exc):
- continue
- raise
- for nested in nested_entries:
- nested_name = nested.get("name", "")
- if not nested_name or nested_name in {".", ".."} or not nested.get("isDir"):
- continue
- nested_path = f"{child_path.rstrip('/')}/{nested_name}"
- if not await self._looks_like_legacy_session_dir(nested_path):
- continue
- if await self._legacy_session_visible(nested_path, ctx, owner_hint=name):
- items.append((nested, f"{output_root_uri}/{nested_name}"))
- return items
-
- async def _session_root_items(
- self,
- uri: str,
- ctx: RequestContext,
- ) -> List[tuple[Dict[str, Any], str]]:
- primary_path = self._uri_to_path(uri, ctx=ctx)
- output_root_uri = uri.rstrip("/")
- owner_user_id = self._safe_uri_parts(uri)[1]
- by_name: Dict[str, tuple[Dict[str, Any], str]] = {}
- try:
- for entry in await self._ls_entries(primary_path, ctx=ctx):
- name = entry.get("name", "")
- if not name or name in {".", ".."}:
- continue
- by_name[name] = (entry, f"{output_root_uri}/{name}")
- except Exception as exc:
- if not is_not_found_error(exc):
- raise
-
- legacy_path = self._legacy_session_path("viking://session", ctx=ctx)
- for entry, entry_uri in await self._legacy_session_root_items(
- legacy_path, ctx, output_root_uri, owner_user_id
- ):
- name = entry.get("name", "")
- if name and name not in by_name:
- by_name[name] = (entry, entry_uri)
- return list(by_name.values())
-
async def _list_read_path_items(
self,
uri: str,
@@ -980,58 +762,37 @@ class _AccessMixin:
ctx: Optional[RequestContext] = None,
) -> tuple[List[tuple[Dict[str, Any], str]], int, bool]:
"""Return one mapped RagFS page, consumed count, and exhaustion state."""
- real_ctx = self._ctx_or_default(ctx)
- if self._is_session_root_uri(uri):
- items = await self._session_root_items(uri, real_ctx)
- return items, len(items), True
-
- primary_path = self._uri_to_path(uri, ctx=ctx)
- merge_paths = self._legacy_session_alias(uri) is not None
- found_path = False
- last_not_found: Optional[Exception] = None
- by_uri: Dict[str, tuple[Dict[str, Any], str]] = {}
- raw_count = 0
-
- for path in self._read_paths(uri, ctx=ctx):
- try:
- entries = await self._ls_entries(
- path,
- offset=0 if merge_paths else raw_offset,
- limit=None if merge_paths else raw_limit,
- sort_by=sort_by,
- sort_order=sort_order,
- filter_internal=False,
- ctx=ctx,
- )
- except Exception as exc:
- if is_not_found_error(exc):
- last_not_found = exc
- continue
- raise
-
- # Missing legacy directories need no owner probes. Check visibility
- # before merging entries from a directory that actually exists.
- if not await self._read_path_visible(uri, path, primary_path, real_ctx):
- continue
+ path = self._uri_to_path(uri, ctx=ctx)
+ try:
+ raw_entries = await self._ls_entries(
+ path,
+ offset=raw_offset,
+ limit=raw_limit,
+ sort_by=sort_by,
+ sort_order=sort_order,
+ filter_internal=False,
+ ctx=ctx,
+ )
+ except Exception as exc:
+ if is_not_found_error(exc):
+ raise NotFoundError(uri, "directory") from exc
+ raise
- found_path = True
- raw_count += len(entries)
- entries = self._filter_ls_entries(path, entries)
- for entry in entries:
- entry_uri = self._alias_uri_for_path(
+ entries = self._filter_ls_entries(path, raw_entries)
+ items = [
+ (
+ entry,
+ self._alias_uri_for_path(
request_uri=uri,
base_path=path,
entry_path=f"{path.rstrip('/')}/{entry.get('name', '')}",
ctx=ctx,
- )
- by_uri.setdefault(entry_uri, (entry, entry_uri))
- if not merge_paths:
- break
-
- if found_path:
- exhausted = raw_limit is None or merge_paths or raw_count < raw_limit
- return list(by_uri.values()), raw_count, exhausted
- raise NotFoundError(uri, "directory") from last_not_found
+ ),
+ )
+ for entry in entries
+ ]
+ exhausted = raw_limit is None or len(raw_entries) < raw_limit
+ return items, len(raw_entries), exhausted
def _path_to_uri(self, path: str, ctx: Optional[RequestContext] = None) -> str:
"""/local/{account}/... -> viking://...
diff --git a/openviking/storage/viking_fs/_grep.py b/openviking/storage/viking_fs/_grep.py
index b9c88fc21..8b60b646e 100644
--- a/openviking/storage/viking_fs/_grep.py
+++ b/openviking/storage/viking_fs/_grep.py
@@ -292,11 +292,7 @@ class _GrepMixin:
after_context=0,
):
"""Filesystem grep path: prefer native agfs grep and fall back if unavailable."""
- native_safe = (
- content_transform is None
- and allowed_uris is None
- and await self._session_native_grep_safe(uri, ctx)
- )
+ native_safe = content_transform is None and allowed_uris is None
if native_safe:
try:
# Session grep historically used the Python fallback, where
@@ -336,37 +332,6 @@ class _GrepMixin:
after_context=after_context,
)
- async def _session_native_grep_safe(self, uri: str, ctx: Optional[RequestContext]) -> bool:
- """Return whether native grep sees every visible path for ``uri``.
-
- Canonical session reads merge the current user namespace with two
- historical storage layouts. Native AGFS grep accepts one physical
- root, so it is complete only when no visible legacy candidate exists.
- """
- legacy_uri = self._legacy_session_alias(uri)
- if legacy_uri is None:
- return True
-
- real_ctx = self._ctx_or_default(ctx)
- if self._is_session_root_uri(uri):
- primary_path = self._uri_to_path(uri, ctx=ctx)
- if not await self._agfs_path_exists(primary_path):
- return False
- legacy_path = self._legacy_session_path(legacy_uri, ctx=ctx)
- owner_user_id = self._safe_uri_parts(uri)[1]
- legacy_items = await self._legacy_session_root_items(
- legacy_path, real_ctx, uri.rstrip("/"), owner_user_id
- )
- return not legacy_items
-
- primary_path = self._uri_to_path(uri, ctx=ctx)
- for path in self._read_paths(uri, ctx=ctx)[1:]:
- if not await self._agfs_path_exists(path):
- continue
- if await self._read_path_visible(uri, path, primary_path, real_ctx):
- return False
- return True
-
async def _grep_vikingdb_then_fs(
self,
uri,
diff --git a/openviking/storage/viking_fs/_ops.py b/openviking/storage/viking_fs/_ops.py
index 2e0e9f8c3..8ddd1fd7b 100644
--- a/openviking/storage/viking_fs/_ops.py
+++ b/openviking/storage/viking_fs/_ops.py
@@ -148,25 +148,17 @@ class _OpsMixin:
real_ctx = self._ctx_or_default(ctx)
uri = await self.resolve_uri(uri, real_ctx)
await self._ensure_access(uri, ctx)
- primary_path = self._uri_to_path(uri, ctx=ctx)
+ path = self._uri_to_path(uri, ctx=ctx)
# Decryption + offset/size slicing now happen inside the ragfs encryption layer
# (when configured); the plaintext stack reads bytes directly. Either way, pass the
# offset/size through and let the Rust layer return the requested slice.
- last_not_found: Optional[Exception] = None
- for path in self._read_paths(uri, ctx=ctx):
- if not await self._read_path_visible(uri, path, primary_path, real_ctx):
- continue
- try:
- result = await self._async_agfs.read(path, offset, size)
- break
- except Exception as exc:
- if is_not_found_error(exc):
- last_not_found = exc
- continue
- raise
- else:
- raise NotFoundError(uri, "file") from last_not_found
+ try:
+ result = await self._async_agfs.read(path, offset, size)
+ except Exception as exc:
+ if is_not_found_error(exc):
+ raise NotFoundError(uri, "file") from exc
+ raise
if isinstance(result, bytes):
raw = result
elif result is not None and hasattr(result, "content"):
@@ -1163,35 +1155,13 @@ class _OpsMixin:
real_ctx = self._ctx_or_default(ctx)
uri = await self.resolve_uri(uri, real_ctx)
await self._ensure_access(uri, ctx)
- primary_path = self._uri_to_path(uri, ctx=ctx)
- path = primary_path
- last_not_found: Optional[Exception] = None
- for candidate_path in self._read_paths(uri, ctx=ctx):
- if not await self._read_path_visible(uri, candidate_path, primary_path, real_ctx):
- continue
- try:
- result = await self._async_agfs.stat(candidate_path)
- path = candidate_path
- break
- except Exception as exc:
- if is_not_found_error(exc):
- last_not_found = exc
- continue
- raise
- else:
- if self._is_session_root_uri(uri):
- now = datetime.now(timezone.utc).isoformat()
- result = {
- "name": "session",
- "size": 0,
- "mode": 0o755,
- "modTime": now,
- "isDir": True,
- }
- if include_lock_status:
- result["isLocked"] = False
- return result
- raise NotFoundError(uri, "file") from last_not_found
+ path = self._uri_to_path(uri, ctx=ctx)
+ try:
+ result = await self._async_agfs.stat(path)
+ except Exception as exc:
+ if is_not_found_error(exc):
+ raise NotFoundError(uri, "file") from exc
+ raise
if isinstance(result, dict):
result["uri"] = uri
if include_lock_status:
@@ -1228,14 +1198,7 @@ class _OpsMixin:
self._safe_uri_parts(uri)
if not self._is_accessible(uri, real_ctx):
return False
-
- primary_path = self._uri_to_path(uri, ctx=ctx)
- for candidate_path in self._read_paths(uri, ctx=ctx):
- if not await self._read_path_visible(uri, candidate_path, primary_path, real_ctx):
- continue
- if await self._agfs_path_exists(candidate_path):
- return True
- return self._is_session_root_uri(uri)
+ return await self._agfs_path_exists(self._uri_to_path(uri, ctx=ctx))
async def glob(
self,
@@ -1259,17 +1222,8 @@ class _OpsMixin:
real_ctx = self._ctx_or_default(ctx)
return_entries = extra_fields is not None
aug_fields = list(extra_fields) if extra_fields else []
- primary_path = self._uri_to_path(uri, ctx=ctx)
- path: Optional[str] = None
- for candidate_path in self._read_paths(uri, ctx=ctx):
- if not await self._read_path_visible(uri, candidate_path, primary_path, real_ctx):
- continue
- if await self._agfs_path_exists(candidate_path):
- path = candidate_path
- break
- if path is None:
- if self._is_session_root_uri(uri):
- return {"matches": [], "count": 0}
+ path = self._uri_to_path(uri, ctx=ctx)
+ if not await self._agfs_path_exists(path):
raise NotFoundError(uri, "directory")
if not force_fs:
@@ -1315,8 +1269,6 @@ class _OpsMixin:
acl_enabled=acl_enabled,
):
continue
- if not await self._read_path_visible(uri, entry["path"], primary_path, real_ctx):
- continue
entry_uri = self._alias_uri_for_path(
request_uri=uri,
base_path=path,
@@ -2080,23 +2032,15 @@ class _OpsMixin:
real_ctx = self._ctx_or_default(ctx)
uri = await self.resolve_uri(uri, real_ctx)
await self._ensure_access(uri, ctx)
- primary_path = self._uri_to_path(uri, ctx=ctx)
+ path = self._uri_to_path(uri, ctx=ctx)
# Verify the file exists before reading, because AGFS read returns
# empty bytes for non-existent files instead of raising an error.
- last_not_found: Optional[Exception] = None
- for path in self._read_paths(uri, ctx=ctx):
- if not await self._read_path_visible(uri, path, primary_path, real_ctx):
- continue
- try:
- stat = await self._async_agfs.stat(path)
- break
- except Exception as exc:
- if is_not_found_error(exc):
- last_not_found = exc
- continue
- raise
- else:
- raise NotFoundError(uri, "file") from last_not_found
+ try:
+ stat = await self._async_agfs.stat(path)
+ except Exception as exc:
+ if is_not_found_error(exc):
+ raise NotFoundError(uri, "file") from exc
+ raise
if isinstance(stat, dict) and stat.get("isDir", False):
raise InvalidArgumentError(
f"Directory URI is not readable as a file: {uri}. "
@@ -2136,21 +2080,13 @@ class _OpsMixin:
real_ctx = self._ctx_or_default(ctx)
uri = await self.resolve_uri(uri, real_ctx)
await self._ensure_access(uri, ctx)
- primary_path = self._uri_to_path(uri, ctx=ctx)
- last_not_found: Optional[Exception] = None
- for path in self._read_paths(uri, ctx=ctx):
- if not await self._read_path_visible(uri, path, primary_path, real_ctx):
- continue
- try:
- stat = await self._async_agfs.stat(path)
- break
- except Exception as exc:
- if is_not_found_error(exc):
- last_not_found = exc
- continue
- raise
- else:
- raise NotFoundError(uri, "file") from last_not_found
+ path = self._uri_to_path(uri, ctx=ctx)
+ try:
+ stat = await self._async_agfs.stat(path)
+ except Exception as exc:
+ if is_not_found_error(exc):
+ raise NotFoundError(uri, "file") from exc
+ raise
if isinstance(stat, dict) and stat.get("isDir", False):
raise InvalidArgumentError(
f"Cannot read directory as file: {uri}",
@@ -2331,67 +2267,6 @@ class _OpsMixin:
overview_limit=overview_limit,
)
- @staticmethod
- def _ls_entry_mtime(entry: Dict[str, Any]) -> Optional[float]:
- raw_time = entry.get("modTime")
- if isinstance(raw_time, (int, float)):
- return float(raw_time)
- if isinstance(raw_time, str) and raw_time:
- try:
- return parse_iso_datetime(raw_time).timestamp()
- except (TypeError, ValueError, OverflowError):
- return None
-
- legacy_time = entry.get("mtime")
- if isinstance(legacy_time, (int, float)):
- return float(legacy_time)
- return None
-
- @classmethod
- def _sort_ls_entry_items(
- cls,
- entry_items: List[tuple[Dict[str, Any], str]],
- sort_by: Optional[str],
- sort_order: str,
- ) -> List[tuple[Dict[str, Any], str]]:
- if sort_by is None:
- return entry_items
-
- descending = sort_order == "desc"
- directories = [item for item in entry_items if item[0].get("isDir", False)]
- files = [item for item in entry_items if not item[0].get("isDir", False)]
-
- def name_key(item: tuple[Dict[str, Any], str]) -> tuple[str, str]:
- """Return the case-insensitive and original entry name."""
- name = str(item[0].get("name", ""))
- return name.lower(), name
-
- if sort_by == "name":
- directories.sort(key=name_key, reverse=descending)
- files.sort(key=name_key, reverse=descending)
- return directories + files
-
- def sort_by_mtime(
- items: List[tuple[Dict[str, Any], str]],
- ) -> List[tuple[Dict[str, Any], str]]:
- timestamped = []
- missing = []
- for item in items:
- timestamp = cls._ls_entry_mtime(item[0])
- if timestamp is None:
- missing.append(item)
- else:
- timestamped.append((timestamp, item))
- timestamped.sort(key=lambda pair: name_key(pair[1]))
- timestamped.sort(
- key=lambda pair: pair[0],
- reverse=descending,
- )
- missing.sort(key=name_key)
- return [item for _, item in timestamped] + missing
-
- return sort_by_mtime(directories) + sort_by_mtime(files)
-
async def _ls_original(
self,
uri: str,
@@ -2473,7 +2348,6 @@ class _OpsMixin:
raw_offset = 0
raw_limit = None if node_limit is None else max(node_limit, 256)
remaining_offset = offset
- merge_paths = self._legacy_session_alias(uri) is not None
browsable: List[tuple[Dict[str, Any], str]] = []
acl_enabled = await self._acl_enabled(ctx)
expose_resource_names = acl_enabled and is_acl_uri(uri)
@@ -2487,8 +2361,6 @@ class _OpsMixin:
sort_order=sort_order,
ctx=ctx,
)
- if merge_paths:
- entry_items = self._sort_ls_entry_items(entry_items, sort_by, sort_order)
entry_items = [
item
for item in entry_items
diff --git a/openviking/storage/viking_fs/_semantic.py b/openviking/storage/viking_fs/_semantic.py
index 2c45873f5..2f31b1905 100644
--- a/openviking/storage/viking_fs/_semantic.py
+++ b/openviking/storage/viking_fs/_semantic.py
@@ -27,7 +27,7 @@ from openviking.storage.viking_fs._base import (
from openviking.telemetry import get_current_telemetry
from openviking.utils.image_search import build_multimodal_embedding_input
from openviking.utils.time_decay import parse_duration_ms
-from openviking_cli.exceptions import InvalidArgumentError, NotFoundError
+from openviking_cli.exceptions import InvalidArgumentError
from openviking_cli.retrieve import ContextType
from openviking_cli.utils.config.retrieval_config import RetrievalConfig
@@ -73,20 +73,7 @@ class _SemanticMixin:
must guarantee that the URI points to a directory.
"""
await self._ensure_access(uri, ctx)
- real_ctx = self._ctx_or_default(ctx)
- primary_path = self._uri_to_path(uri, ctx=ctx)
- for path in self._read_paths(uri, ctx=ctx):
- if not await self._read_path_visible(uri, path, primary_path, real_ctx):
- continue
- try:
- if not await self._agfs_path_exists(path):
- continue
- return await self._read_abstract_file(path, uri, ctx=ctx)
- except Exception as exc:
- if is_not_found_error(exc):
- continue
- raise
- return f"# {uri} [Directory abstract is not ready]"
+ return await self._read_abstract_file(self._uri_to_path(uri, ctx=ctx), uri, ctx=ctx)
async def abstract(
self,
@@ -99,31 +86,14 @@ class _SemanticMixin:
the endpoint remains usable for both file and directory URIs.
"""
await self._ensure_access(uri, ctx)
- real_ctx = self._ctx_or_default(ctx)
- primary_path = self._uri_to_path(uri, ctx=ctx)
- path = primary_path
- last_exc: Optional[Exception] = None
- for candidate_path in self._read_paths(uri, ctx=ctx):
- if not await self._read_path_visible(uri, candidate_path, primary_path, real_ctx):
- continue
- try:
- info = await self._async_agfs.stat(candidate_path)
- path = candidate_path
- break
- except Exception as exc:
- if is_not_found_error(exc):
- last_exc = exc
- continue
- mapped = map_exception(exc, resource=uri)
- if mapped is not None:
- raise mapped from exc
- raise
- else:
- if last_exc is not None:
- mapped = map_exception(last_exc, resource=uri)
- if mapped is not None:
- raise mapped from last_exc
- raise NotFoundError(uri, "directory") from last_exc
+ path = self._uri_to_path(uri, ctx=ctx)
+ try:
+ info = await self._async_agfs.stat(path)
+ except Exception as exc:
+ mapped = map_exception(exc, resource=uri)
+ if mapped is not None:
+ raise mapped from exc
+ raise
if not info.get("isDir", info.get("is_dir")):
parent_path = path.rsplit("/", 1)[0] or "/"
parent_uri = self._path_to_uri(parent_path, ctx=ctx)
@@ -146,31 +116,14 @@ class _SemanticMixin:
the endpoint remains usable for both file and directory URIs.
"""
await self._ensure_access(uri, ctx)
- real_ctx = self._ctx_or_default(ctx)
- primary_path = self._uri_to_path(uri, ctx=ctx)
- path = primary_path
- last_exc: Optional[Exception] = None
- for candidate_path in self._read_paths(uri, ctx=ctx):
- if not await self._read_path_visible(uri, candidate_path, primary_path, real_ctx):
- continue
- try:
- info = await self._async_agfs.stat(candidate_path)
- path = candidate_path
- break
- except Exception as exc:
- if is_not_found_error(exc):
- last_exc = exc
- continue
- mapped = map_exception(exc, resource=uri)
- if mapped is not None:
- raise mapped from exc
- raise
- else:
- if last_exc is not None:
- mapped = map_exception(last_exc, resource=uri)
- if mapped is not None:
- raise mapped from last_exc
- raise NotFoundError(uri, "directory") from last_exc
+ path = self._uri_to_path(uri, ctx=ctx)
+ try:
+ info = await self._async_agfs.stat(path)
+ except Exception as exc:
+ mapped = map_exception(exc, resource=uri)
+ if mapped is not None:
+ raise mapped from exc
+ raise
if not info.get("isDir", info.get("is_dir")):
parent_path = path.rsplit("/", 1)[0] or "/"
parent_uri = self._path_to_uri(parent_path, ctx=ctx)
diff --git a/openviking_cli/utils/uri.py b/openviking_cli/utils/uri.py
index 6a7123b52..853ef75a7 100644
--- a/openviking_cli/utils/uri.py
+++ b/openviking_cli/utils/uri.py
@@ -41,7 +41,7 @@ class VikingURI:
- resources: Independent resource scope (viking://resources/{project}/...)
- user: Explicit user scope (viking://user/{user_id}/...), including sessions under
viking://user/{user_id}/sessions/{session_id}.
- - session: Legacy alias for user sessions (viking://session/{session_id}/...)
+ - session: Historical scope parsed only for explicit data migration
- agent: Agent capabilities scope (viking://agent/skills/..., viking://agent/endpoints/...)
- system: Account-level system assets (viking://system/...)
- queue: Queue scope (viking://queue/...)
diff --git a/sdk/go/client_test.go b/sdk/go/client_test.go
index cf710ff28..d9124442a 100644
--- a/sdk/go/client_test.go
+++ b/sdk/go/client_test.go
@@ -351,7 +351,7 @@ func TestListAndTreePreserveOptionsAndPagination(t *testing.T) {
}))
defer closeServer()
- listPage, err := client.ListPage(context.Background(), "viking://session", &ListOptions{
+ listPage, err := client.ListPage(context.Background(), "viking://~/sessions", &ListOptions{
NodeLimit: 200,
Offset: 4,
Limit: 5,
@@ -370,9 +370,9 @@ func TestListAndTreePreserveOptionsAndPagination(t *testing.T) {
if !listPage.HasMore || !reflect.DeepEqual(listPage.Result, entries) {
t.Fatalf("list page = %#v", listPage)
}
- legacyList, err := client.List(context.Background(), "viking://session", nil)
- if err != nil || !reflect.DeepEqual(legacyList, entries) {
- t.Fatalf("list = %#v, err = %v", legacyList, err)
+ sessionList, err := client.List(context.Background(), "viking://~/sessions", nil)
+ if err != nil || !reflect.DeepEqual(sessionList, entries) {
+ t.Fatalf("list = %#v, err = %v", sessionList, err)
}
treePage, err := client.TreePage(context.Background(), "viking://resources/docs", &TreeOptions{
NodeLimit: 200,
diff --git a/sdk/typescript/tests/client.test.ts b/sdk/typescript/tests/client.test.ts
index 47d6da874..c5d1453d4 100644
--- a/sdk/typescript/tests/client.test.ts
+++ b/sdk/typescript/tests/client.test.ts
@@ -659,7 +659,7 @@ describe("OpenVikingClient", () => {
fetch: fetcher,
});
- const listPage = await client.listPage("viking://session", {
+ const listPage = await client.listPage("viking://~/sessions", {
nodeLimit: 200,
offset: 4,
limit: 5,
@@ -671,7 +671,7 @@ describe("OpenVikingClient", () => {
extraFields: ["locked", "id"],
});
expect(listPage).toEqual({ result: entries, hasMore: true });
- await expect(client.list("viking://session")).resolves.toEqual(entries);
+ await expect(client.list("viking://~/sessions")).resolves.toEqual(entries);
const treePage = await client.treePage("viking://resources/docs", {
levelLimit: 2,
offset: 6,
diff --git a/tests/server/test_api_sessions.py b/tests/server/test_api_sessions.py
index 769e9347f..f2db21aa9 100644
--- a/tests/server/test_api_sessions.py
+++ b/tests/server/test_api_sessions.py
@@ -170,23 +170,6 @@ async def test_get_session(client: httpx.AsyncClient):
assert body["result"]["uri"] == session_uri
-async def test_legacy_session_uri_alias_reads_current_user_session(client: httpx.AsyncClient):
- session_id = "legacy-alias-read"
- await client.post("/api/v1/sessions", json={"session_id": session_id})
- await client.post(
- f"/api/v1/sessions/{session_id}/messages",
- json=_message_request("user", content="legacy alias message"),
- )
-
- resp = await client.get(
- "/api/v1/content/read",
- params={"uri": f"viking://session/{session_id}/messages.jsonl"},
- )
-
- assert resp.status_code == 200
- assert "legacy alias message" in resp.json()["result"]
-
-
@pytest.mark.parametrize("artifact", ["empty", "recall_ledger", "meta"])
async def test_add_message_repairs_session_directory_without_live_messages(
client: httpx.AsyncClient,
diff --git a/tests/service/test_session_service_metrics.py b/tests/service/test_session_service_metrics.py
index 6b63df44b..0f14e8929 100644
--- a/tests/service/test_session_service_metrics.py
+++ b/tests/service/test_session_service_metrics.py
@@ -87,7 +87,7 @@ async def test_sessions_returns_empty_and_logs_when_storage_listing_fails(
@pytest.mark.asyncio
-async def test_sessions_uses_canonical_scope_and_relies_on_storage_compatibility():
+async def test_sessions_uses_canonical_scope():
service = SessionService(viking_fs=Mock())
ctx = _make_ctx()
@@ -97,11 +97,6 @@ async def test_sessions_uses_canonical_scope_and_relies_on_storage_compatibility
return [
{"name": "duplicate", "isDir": True, "modTime": "2026-07-13T01:00:00Z"},
{"name": "new-session", "isDir": True, "modTime": "2026-07-13T02:00:00Z"},
- {
- "name": "legacy-session",
- "isDir": True,
- "modTime": "2026-07-12T01:00:00Z",
- },
]
service._viking_fs.ls = AsyncMock(side_effect=_ls)
@@ -121,12 +116,6 @@ async def test_sessions_uses_canonical_scope_and_relies_on_storage_compatibility
"is_dir": True,
"mod_time": "2026-07-13T02:00:00Z",
},
- {
- "session_id": "legacy-session",
- "uri": "viking://user/alice/sessions/legacy-session",
- "is_dir": True,
- "mod_time": "2026-07-12T01:00:00Z",
- },
]
service._viking_fs.ls.assert_awaited_once()
diff --git a/tests/session/memory/test_agent_experience_context_provider.py b/tests/session/memory/test_agent_experience_context_provider.py
index 1a0660c66..c1a944919 100644
--- a/tests/session/memory/test_agent_experience_context_provider.py
+++ b/tests/session/memory/test_agent_experience_context_provider.py
@@ -140,7 +140,7 @@ async def test_agent_experience_prefetch_includes_structured_read_results():
"task_query": "Discuss the album release party.",
"retrieval_anchor": "Stage: final_response; Boundary: completed discussion.",
"case_name": "discussion_case",
- "source_archive_uri": "viking://session/previous/archive/001",
+ "source_archive_uri": "viking://user/default/sessions/previous/archive/001",
"last_update_trace_id": "source-trace-id",
"version": 3,
},
diff --git a/tests/session/memory/test_memory_react_system_prompt.py b/tests/session/memory/test_memory_react_system_prompt.py
index 0bbeee9a7..42635eef2 100644
--- a/tests/session/memory/test_memory_react_system_prompt.py
+++ b/tests/session/memory/test_memory_react_system_prompt.py
@@ -146,7 +146,7 @@ class TestSessionConversationToolFiltering:
ToolPart(
tool_id="tool_1",
tool_name="read",
- tool_uri="viking://session/test/tools/tool_1",
+ tool_uri="viking://user/default/sessions/test/tools/tool_1",
skill_uri="viking://user/skills/create_presentation",
tool_input={"file_path": "/skills/ppt/SKILL.md"},
tool_output="ok",
@@ -174,7 +174,7 @@ class TestSessionConversationToolFiltering:
ToolPart(
tool_id="tool_1",
tool_name="read",
- tool_uri="viking://session/test/tools/tool_1",
+ tool_uri="viking://user/default/sessions/test/tools/tool_1",
tool_input={"file_path": "README.md"},
tool_output="ok",
tool_status="completed",
@@ -250,7 +250,7 @@ class TestSessionConversationToolFiltering:
ToolPart(
tool_id="tool_1",
tool_name="read",
- tool_uri="viking://session/test/tools/tool_1",
+ tool_uri="viking://user/default/sessions/test/tools/tool_1",
tool_input={"file_path": "README.md"},
tool_output="这是中文工具输出",
tool_status="completed",
diff --git a/tests/session/test_tool_result_synopsis.py b/tests/session/test_tool_result_synopsis.py
index 3dcf24f4c..9de2fe64d 100644
--- a/tests/session/test_tool_result_synopsis.py
+++ b/tests/session/test_tool_result_synopsis.py
@@ -27,7 +27,7 @@ def test_render_stub_preserves_navigation_contract():
rendered = render_tool_result_stub(
synopsis,
- ref="viking://session/s1/tool-results/tr_123",
+ ref="viking://user/default/sessions/s1/tool-results/tr_123",
tool_name="read_file",
sha256="0123456789abcdef0123",
reason="single_threshold",
@@ -40,7 +40,7 @@ def test_render_stub_preserves_navigation_contract():
assert "tool_name: read_file" in rendered
assert "original_chars: 120" in rendered
assert "preview_chars: 80" in rendered
- assert "ref: viking://session/s1/tool-results/tr_123" in rendered
+ assert "ref: viking://user/default/sessions/s1/tool-results/tr_123" in rendered
assert "openviking_tool_result_search" in rendered
assert "openviking_tool_result_read" in rendered
diff --git a/tests/storage/test_session_delete_storage.py b/tests/storage/test_session_delete_storage.py
index ce4e5b79e..77cdcb120 100644
--- a/tests/storage/test_session_delete_storage.py
+++ b/tests/storage/test_session_delete_storage.py
@@ -48,7 +48,6 @@ def storage():
"request_uri",
[
SESSION_URI,
- "viking://session/session-1",
"viking://~/sessions/session-1",
"viking://temp/alice/raw-upload-1",
],
diff --git a/tests/storage/test_viking_fs_actor_peer_view.py b/tests/storage/test_viking_fs_actor_peer_view.py
index 9513fa900..5d22e1b9d 100644
--- a/tests/storage/test_viking_fs_actor_peer_view.py
+++ b/tests/storage/test_viking_fs_actor_peer_view.py
@@ -24,28 +24,6 @@ class _MemoryAGFS:
b"zhang"
),
"/local/acct/user/support_bot/resources/guide.md": b"guide",
- "/local/acct/user/support_bot/sessions/duplicate/messages.jsonl": (
- b'{"role":"user","content":"new"}\n'
- ),
- "/local/acct/user/support_bot/sessions/new-session/messages.jsonl": (
- b'{"role":"user","content":"new only"}\n'
- ),
- "/local/acct/session/duplicate/messages.jsonl": (
- b'{"role":"user","content":"legacy duplicate"}\n'
- ),
- "/local/acct/session/legacy-session/messages.jsonl": (
- b'{"role":"user","content":"legacy"}\n'
- ),
- "/local/acct/session/legacy-session/.meta.json": (
- b'{"created_by_user_id":"support_bot"}'
- ),
- "/local/acct/session/other-owned/.meta.json": b'{"created_by_user_id":"other"}',
- "/local/acct/session/other-owned/messages.jsonl": (
- b'{"role":"user","content":"other"}\n'
- ),
- "/local/acct/session/support_bot/nested-session/messages.jsonl": (
- b'{"role":"user","content":"nested"}\n'
- ),
}
self.dirs = set()
for path in self.files:
@@ -211,100 +189,6 @@ async def test_agent_directories_are_shared_within_account(fs, actor_ctx):
await fs.read_file("viking://agent/skills/demo/SKILL.md", ctx=other_account)
-async def test_legacy_session_scope_merges_new_and_unmigrated_sessions(fs, actor_ctx):
- session_root = "viking://user/support_bot/sessions"
- entries = await fs.ls(session_root, ctx=actor_ctx)
-
- assert [entry["uri"] for entry in entries] == [
- f"{session_root}/duplicate",
- f"{session_root}/new-session",
- f"{session_root}/legacy-session",
- f"{session_root}/nested-session",
- ]
-
- assert (
- await fs.read_file(f"{session_root}/duplicate/messages.jsonl", ctx=actor_ctx)
- == '{"role":"user","content":"new"}\n'
- )
- assert (
- await fs.read_file(f"{session_root}/legacy-session/messages.jsonl", ctx=actor_ctx)
- == '{"role":"user","content":"legacy"}\n'
- )
- assert (
- await fs.read_file(f"{session_root}/nested-session/messages.jsonl", ctx=actor_ctx)
- == '{"role":"user","content":"nested"}\n'
- )
- with pytest.raises(NotFoundError):
- await fs.read_file(f"{session_root}/other-owned/messages.jsonl", ctx=actor_ctx)
-
- for session_id in ("new-session", "legacy-session", "nested-session"):
- children = await fs.ls(f"{session_root}/{session_id}", ctx=actor_ctx)
- assert [entry["uri"] for entry in children] == [
- f"{session_root}/{session_id}/messages.jsonl"
- ]
- with pytest.raises(NotFoundError):
- await fs.ls(f"{session_root}/other-owned", ctx=actor_ctx)
-
-
-async def test_session_grep_preserves_legacy_merge_and_primary_shadow(fs, actor_ctx):
- session_root = "viking://user/support_bot/sessions"
-
- result = await fs.grep(
- session_root,
- pattern="new|legacy|nested|other",
- ctx=actor_ctx,
- )
-
- assert result["matches"] == [
- {
- "uri": f"{session_root}/duplicate/messages.jsonl",
- "line": 1,
- "content": '{"role":"user","content":"new"}',
- },
- {
- "uri": f"{session_root}/new-session/messages.jsonl",
- "line": 1,
- "content": '{"role":"user","content":"new only"}',
- },
- {
- "uri": f"{session_root}/legacy-session/messages.jsonl",
- "line": 1,
- "content": '{"role":"user","content":"legacy"}',
- },
- {
- "uri": f"{session_root}/nested-session/messages.jsonl",
- "line": 1,
- "content": '{"role":"user","content":"nested"}',
- },
- ]
-
-
-async def test_session_native_grep_gate_allows_primary_only_session(fs, actor_ctx):
- assert await fs._session_native_grep_safe(
- "viking://user/support_bot/sessions/new-session", actor_ctx
- )
-
-
-async def test_session_native_grep_gate_rejects_visible_legacy_layouts(fs, actor_ctx):
- assert not await fs._session_native_grep_safe(
- "viking://user/support_bot/sessions/legacy-session", actor_ctx
- )
- assert not await fs._session_native_grep_safe(
- "viking://user/support_bot/sessions/nested-session", actor_ctx
- )
-
-
-async def test_session_native_grep_gate_ignores_other_owner_legacy_data(fs, actor_ctx):
- assert await fs._session_native_grep_safe(
- "viking://user/support_bot/sessions/other-owned", actor_ctx
- )
-
-
@pytest.mark.asyncio
async def test_actor_peer_view_filters_tree_from_user_root(fs, actor_ctx):
entries = await fs.tree(
diff --git a/tests/storage/test_viking_fs_glob.py b/tests/storage/test_viking_fs_glob.py
index c651a5739..fba882981 100644
--- a/tests/storage/test_viking_fs_glob.py
+++ b/tests/storage/test_viking_fs_glob.py
@@ -88,11 +88,7 @@ def fs(monkeypatch):
monkeypatch.setattr(
viking_fs, "_uri_to_path", lambda _uri, **_kwargs: "/local/test_account/resources"
)
- monkeypatch.setattr(
- viking_fs, "_read_paths", lambda _uri, **_kwargs: ["/local/test_account/resources"]
- )
monkeypatch.setattr(viking_fs, "_agfs_path_exists", lambda _path: _async_true())
- monkeypatch.setattr(viking_fs, "_read_path_visible", lambda *_args, **_kwargs: _async_true())
monkeypatch.setattr(
viking_fs,
"_path_to_uri",
@@ -816,7 +812,6 @@ async def test_glob_checks_access_before_listing(monkeypatch, fs):
@pytest.mark.asyncio
async def test_glob_preserves_request_uri_alias(monkeypatch, fs):
monkeypatch.setattr(fs, "_uri_to_path", lambda _uri, **_kwargs: "/local/test_account/user")
- monkeypatch.setattr(fs, "_read_paths", lambda _uri, **_kwargs: ["/local/test_account/user"])
monkeypatch.setattr(
fs, "_path_to_uri", lambda _path, **_kwargs: "viking://user/test_account/demo.md"
)
@@ -844,7 +839,6 @@ async def test_glob_preserves_request_uri_alias(monkeypatch, fs):
@pytest.mark.asyncio
async def test_glob_preserves_root_uri(monkeypatch, fs):
monkeypatch.setattr(fs, "_uri_to_path", lambda _uri, **_kwargs: "/local/test_account")
- monkeypatch.setattr(fs, "_read_paths", lambda _uri, **_kwargs: ["/local/test_account"])
monkeypatch.setattr(
fs, "_path_to_uri", lambda _path, **_kwargs: "viking://resources/should-not-leak.md"
)
@@ -975,11 +969,6 @@ async def test_glob_preserves_canonical_session_uri(monkeypatch, fs):
monkeypatch.setattr(
fs, "_uri_to_path", lambda _uri, **_kwargs: "/local/test_account/user/alice/sessions/sess_1"
)
- monkeypatch.setattr(
- fs,
- "_read_paths",
- lambda _uri, **_kwargs: ["/local/test_account/user/alice/sessions/sess_1"],
- )
async def fake_glob_directory(path, pattern, **kwargs):
return {
@@ -1009,18 +998,13 @@ async def test_glob_preserves_canonical_session_uri(monkeypatch, fs):
@pytest.mark.asyncio
-async def test_glob_uses_path_to_uri_for_non_legacy_namespace(monkeypatch, fs):
- """中文注释:非 legacy 命名空间必须回落到 _path_to_uri,避免错误沿用请求别名。"""
+async def test_glob_uses_path_to_uri_when_request_mapping_differs(monkeypatch, fs):
+ """Return the physical path's URI when the requested mapping differs."""
monkeypatch.setattr(
fs,
"_uri_to_path",
lambda _uri, **_kwargs: "/local/test_account/resources/actual-root",
)
- monkeypatch.setattr(
- fs,
- "_read_paths",
- lambda _uri, **_kwargs: ["/local/test_account/resources/actual-root"],
- )
monkeypatch.setattr(
fs,
"_path_to_uri",
diff --git a/tests/storage/test_viking_fs_grep.py b/tests/storage/test_viking_fs_grep.py
index 06be08078..5769f4625 100644
--- a/tests/storage/test_viking_fs_grep.py
+++ b/tests/storage/test_viking_fs_grep.py
@@ -424,14 +424,7 @@ async def test_grep_without_config_uses_documented_remote_threshold(monkeypatch)
@pytest.mark.asyncio
- "uri",
- [
- "viking://user/alice/sessions/session-1",
- "viking://session/session-1",
- ],
-)
-async def test_session_grep_forces_fs_engine_in_auto_mode(monkeypatch, uri):
+async def test_session_grep_forces_fs_engine_in_auto_mode(monkeypatch):
viking_fs = VikingFS(agfs=_DummyAgfs())
monkeypatch.setattr(viking_fs, "stat", AsyncMock(return_value={"isDir": True}))
resolve_engine = AsyncMock(return_value="vikingdb_then_fs")
@@ -444,7 +437,7 @@ async def test_session_grep_forces_fs_engine_in_auto_mode(monkeypatch, uri):
monkeypatch.setattr(viking_fs, "_grep_vikingdb_then_fs", grep_vikingdb)
await viking_fs.grep(
- uri,
+ "viking://user/alice/sessions/session-1",
pattern="needle",
)
@@ -454,12 +447,11 @@ async def test_session_grep_forces_fs_engine_in_auto_mode(monkeypatch, uri):
@pytest.mark.asyncio
-async def test_primary_only_session_grep_uses_native_agfs(monkeypatch):
+async def test_session_grep_uses_native_agfs(monkeypatch):
viking_fs = VikingFS(agfs=_DummyAgfs())
native_result = {"matches": [], "count": 0, "match_count": 0, "files_scanned": 4}
native_grep = AsyncMock(return_value=native_result)
fallback_grep = AsyncMock()
- monkeypatch.setattr(viking_fs, "_session_native_grep_safe", AsyncMock(return_value=True))
monkeypatch.setattr(viking_fs, "_grep_with_agfs", native_grep)
monkeypatch.setattr(viking_fs, "_grep_encrypted", fallback_grep)
@@ -488,92 +480,6 @@ async def test_primary_only_session_grep_uses_native_agfs(monkeypatch):
fallback_grep.assert_not_awaited()
-async def test_session_grep_with_visible_legacy_data_uses_merge_fallback(monkeypatch):
- viking_fs = VikingFS(agfs=_DummyAgfs())
- fallback_result = {"matches": [], "count": 0, "match_count": 0, "files_scanned": 2}
- native_grep = AsyncMock()
- fallback_grep = AsyncMock(return_value=fallback_result)
- monkeypatch.setattr(viking_fs, "_session_native_grep_safe", AsyncMock(return_value=False))
- monkeypatch.setattr(viking_fs, "_grep_with_agfs", native_grep)
- monkeypatch.setattr(viking_fs, "_grep_encrypted", fallback_grep)
-
- result = await viking_fs._grep_fs(
- uri="viking://user/alice/sessions/session-1",
- pattern="needle",
- exclude_uri=None,
- case_insensitive=False,
- node_limit=None,
- level_limit=10,
- ctx=None,
- )
-
- assert result == fallback_result
- native_grep.assert_not_awaited()
- fallback_grep.assert_awaited_once()
- assert fallback_grep.await_args.kwargs["level_limit"] == 10
-
-
-async def test_session_native_and_legacy_fallback_keep_level_limit_results_equal(monkeypatch):
- viking_fs = VikingFS(agfs=_DummyAgfs())
- match = {
- "uri": "viking://user/alice/sessions/s1/messages.jsonl",
- "line": 1,
- "content": "payment failed",
- }
-
- async def fake_native_grep(**kwargs):
- matches = [match] if kwargs["level_limit"] >= 2 else []
- return {
- "matches": matches,
- "count": len(matches),
- "match_count": len(matches),
- "files_scanned": len(matches),
- }
-
- async def fake_fallback_grep(**kwargs):
- matches = [match] if kwargs["level_limit"] >= 1 else []
- return {
- "matches": matches,
- "count": len(matches),
- "match_count": len(matches),
- "files_scanned": len(matches),
- }
-
- monkeypatch.setattr(viking_fs, "_grep_with_agfs", fake_native_grep)
- monkeypatch.setattr(viking_fs, "_grep_encrypted", fake_fallback_grep)
- native_safe = AsyncMock(side_effect=[True, False])
- monkeypatch.setattr(viking_fs, "_session_native_grep_safe", native_safe)
- kwargs = {
- "uri": "viking://user/alice/sessions",
- "pattern": "payment failed",
- "exclude_uri": None,
- "case_insensitive": False,
- "node_limit": None,
- "level_limit": 1,
- "ctx": None,
- }
-
- assert await viking_fs._grep_fs(**kwargs) == await viking_fs._grep_fs(**kwargs)
-
-
-async def test_virtual_empty_session_root_uses_merge_fallback(monkeypatch):
- viking_fs = VikingFS(agfs=_DummyAgfs())
- primary_path = "/local/default/user/alice/sessions"
- path_exists = AsyncMock(return_value=False)
- legacy_items = AsyncMock(return_value=[])
- monkeypatch.setattr(viking_fs, "_agfs_path_exists", path_exists)
- monkeypatch.setattr(viking_fs, "_legacy_session_root_items", legacy_items)
-
- assert not await viking_fs._session_native_grep_safe(
- "viking://user/alice/sessions", None
- )
- path_exists.assert_awaited_once_with(primary_path)
- legacy_items.assert_not_awaited()
-
-
@pytest.mark.asyncio
@pytest.mark.parametrize(
("node_limit", "expected_remote_limit"),
diff --git a/tests/storage/test_viking_fs_tree.py b/tests/storage/test_viking_fs_tree.py
index 8130cb862..c2d89f83b 100644
--- a/tests/storage/test_viking_fs_tree.py
+++ b/tests/storage/test_viking_fs_tree.py
@@ -40,9 +40,6 @@ async def test_stat_queries_lock_status_only_when_requested(monkeypatch, fs):
async def ensure_access(_uri, _ctx):
return None
- async def read_path_visible(_uri, _path, _primary_path, _ctx):
- return True
-
async def stat_path(_path):
return {"name": "example.md", "isDir": False}
@@ -53,8 +50,6 @@ async def test_stat_queries_lock_status_only_when_requested(monkeypatch, fs):
monkeypatch.setattr(fs, "resolve_uri", resolve_uri)
monkeypatch.setattr(fs, "_ensure_access", ensure_access)
monkeypatch.setattr(fs, "_uri_to_path", lambda _uri, **_kwargs: path)
- monkeypatch.setattr(fs, "_read_paths", lambda _uri, **_kwargs: [path])
- monkeypatch.setattr(fs, "_read_path_visible", read_path_visible)
monkeypatch.setattr(fs._async_agfs, "stat", stat_path)
monkeypatch.setattr(fs, "_is_path_locked_async", is_path_locked)
@@ -809,16 +804,6 @@ async def test_ls_agent_preserves_metadata_acl_and_summary_controls(monkeypatch,
abstract.assert_not_awaited()
overview.assert_awaited_once_with("viking://resources/docs", ctx=_default_ctx())
- tied = [
- ({"name": name, "isDir": False, "modTime": "2026-01-01T00:00:00Z"}, name)
- for name in ["b.md", "a.md"]
- ]
- assert [entry["name"] for entry, _ in fs._sort_ls_entry_items(tied, "mtime", "desc")] == [
- "a.md",
- "b.md",
- ]
-
-
@pytest.mark.asyncio
async def test_tree_filters_and_paginates_before_summary_reads(monkeypatch, fs):
"""Only selected directories consume node slots and summary reads."""
diff --git a/tests/test_token_estimation.py b/tests/test_token_estimation.py
index f9749fac7..0333eefaa 100644
--- a/tests/test_token_estimation.py
+++ b/tests/test_token_estimation.py
@@ -36,10 +36,10 @@ async def test_archive_overview_tokens_do_not_trust_stale_low_metadata():
del uri, ctx
return '{"overview_tokens": 1}'
- store = ArchiveStore(FakeFS(), None, "viking://session/test")
+ store = ArchiveStore(FakeFS(), None, "viking://user/default/sessions/test")
tokens = await store.read_overview_tokens(
- "viking://session/test/history/archive_001",
+ "viking://user/default/sessions/test/history/archive_001",
"\u4f60\u597d\u4e16\u754c",
)
@@ -52,10 +52,10 @@ async def test_archive_overview_tokens_keep_higher_metadata_estimate():
del uri, ctx
return '{"overview_tokens": 5}'
- store = ArchiveStore(FakeFS(), None, "viking://session/test")
+ store = ArchiveStore(FakeFS(), None, "viking://user/default/sessions/test")
tokens = await store.read_overview_tokens(
- "viking://session/test/history/archive_001",
+ "viking://user/default/sessions/test/history/archive_001",
"abcd",
)
diff --git a/tests/unit/test_message.py b/tests/unit/test_message.py
index 77352d8b3..195471862 100644
--- a/tests/unit/test_message.py
+++ b/tests/unit/test_message.py
@@ -160,7 +160,7 @@ class TestToolPart:
part = ToolPart(
tool_id="call-123",
tool_name="search",
- tool_uri="viking://session/test/tools/call-123",
+ tool_uri="viking://user/default/sessions/test/tools/call-123",
skill_uri="viking://user/test/skills/search",
tool_input={"query": "test"},
tool_output="Result",
@@ -168,7 +168,7 @@ class TestToolPart:
duration_ms=150.5,
prompt_tokens=100,
completion_tokens=50,
- tool_output_ref="viking://session/s1/tool-results/tr_call",
+ tool_output_ref="viking://user/default/sessions/s1/tool-results/tr_call",
tool_output_truncated=True,
tool_output_original_chars=1000,
tool_output_preview_chars=100,
@@ -179,7 +179,7 @@ class TestToolPart:
assert part.tool_id == "call-123"
assert part.tool_name == "search"
- assert part.tool_uri == "viking://session/test/tools/call-123"
+ assert part.tool_uri == "viking://user/default/sessions/test/tools/call-123"
assert part.skill_uri == "viking://user/test/skills/search"
assert part.tool_input == {"query": "test"}
assert part.tool_output == "Result"
@@ -187,7 +187,7 @@ class TestToolPart:
assert part.duration_ms == 150.5
assert part.prompt_tokens == 100
assert part.completion_tokens == 50
- assert part.tool_output_ref == "viking://session/s1/tool-results/tr_call"
+ assert part.tool_output_ref == "viking://user/default/sessions/s1/tool-results/tr_call"
assert part.tool_output_truncated is True
assert part.tool_output_original_chars == 1000
assert part.tool_output_preview_chars == 100
@@ -256,7 +256,7 @@ class TestPartFromDict:
"type": "tool",
"tool_id": "call-123",
"tool_name": "search",
- "tool_uri": "viking://session/test/tools/call-123",
+ "tool_uri": "viking://user/default/sessions/test/tools/call-123",
"skill_uri": "viking://user/test/skills/search",
"tool_input": {"query": "test"},
"tool_output": "Result",
@@ -264,7 +264,7 @@ class TestPartFromDict:
"duration_ms": 150.0,
"prompt_tokens": 100,
"completion_tokens": 50,
- "tool_output_ref": "viking://session/s1/tool-results/tr_call",
+ "tool_output_ref": "viking://user/default/sessions/s1/tool-results/tr_call",
"tool_output_truncated": True,
"tool_output_original_chars": 1000,
"tool_output_preview_chars": 100,
@@ -277,7 +277,7 @@ class TestPartFromDict:
assert part.tool_id == "call-123"
assert part.tool_name == "search"
assert part.tool_status == "completed"
- assert part.tool_output_ref == "viking://session/s1/tool-results/tr_call"
+ assert part.tool_output_ref == "viking://user/default/sessions/s1/tool-results/tr_call"
assert part.tool_output_truncated is True
assert part.tool_output_original_chars == 1000
assert part.tool_output_preview_chars == 100
@@ -525,7 +525,7 @@ class TestMessageToDict:
duration_ms=100,
prompt_tokens=50,
completion_tokens=25,
- tool_output_ref="viking://session/s1/tool-results/tr_call",
+ tool_output_ref="viking://user/default/sessions/s1/tool-results/tr_call",
tool_output_truncated=True,
tool_output_original_chars=1000,
tool_output_preview_chars=100,
@@ -539,7 +539,7 @@ class TestMessageToDict:
assert d["parts"][0]["type"] == "tool"
assert d["parts"][0]["tool_id"] == "call-1"
assert d["parts"][0]["duration_ms"] == 100
- assert d["parts"][0]["tool_output_ref"] == "viking://session/s1/tool-results/tr_call"
+ assert d["parts"][0]["tool_output_ref"] == "viking://user/default/sessions/s1/tool-results/tr_call"
assert d["parts"][0]["tool_output_truncated"] is True
assert d["parts"][0]["tool_output_original_chars"] == 1000
assert d["parts"][0]["tool_output_preview_chars"] == 100
@@ -677,7 +677,7 @@ class TestMessageFromDict:
"tool_name": "search",
"tool_uri": "viking://tools/1",
"tool_status": "completed",
- "tool_output_ref": "viking://session/s1/tool-results/tr_call",
+ "tool_output_ref": "viking://user/default/sessions/s1/tool-results/tr_call",
"tool_output_truncated": True,
"tool_output_original_chars": 1000,
"tool_output_preview_chars": 100,
@@ -691,7 +691,7 @@ class TestMessageFromDict:
assert isinstance(msg.parts[0], ToolPart)
assert msg.parts[0].tool_id == "call-1"
- assert msg.parts[0].tool_output_ref == "viking://session/s1/tool-results/tr_call"
+ assert msg.parts[0].tool_output_ref == "viking://user/default/sessions/s1/tool-results/tr_call"
assert msg.parts[0].tool_output_truncated is True
assert msg.parts[0].tool_output_original_chars == 1000
assert msg.parts[0].tool_output_preview_chars == 100
diff --git a/tests/unit/test_namespace_uri_classification.py b/tests/unit/test_namespace_uri_classification.py
index 5a9fbe831..6ddbca079 100644
--- a/tests/unit/test_namespace_uri_classification.py
+++ b/tests/unit/test_namespace_uri_classification.py
@@ -143,21 +143,10 @@ def test_session_uri_helpers_use_user_namespace():
# hint pointing at the '~' home alias.
with pytest.raises(NamespaceShapeError, match=re.escape("viking://~/sessions/s1")):
resolve_request_uri("viking://user/sessions/s1", ctx)
- assert (
- resolve_request_uri("viking://session/s1", ctx)
- == "viking://user/alice/sessions/s1"
- )
- assert (
- resolve_request_uri(
- "viking://session/s1/history/archive_001/messages.jsonl", ctx
- )
- == "viking://user/alice/sessions/s1/history/archive_001/messages.jsonl"
- )
assert is_session_uri("viking://user/alice/sessions/s1")
assert not is_session_uri("viking://user/sessions/s1")
- assert is_session_uri("viking://session/s1")
roots = visible_roots(ctx)
- assert "viking://session" not in roots
+ assert "viking://user/alice" in roots
assert "viking://agent" in roots
diff --git a/tests/unit/test_summarizer_context_type.py b/tests/unit/test_summarizer_context_type.py
index 1b5d6a378..3493cefbc 100644
--- a/tests/unit/test_summarizer_context_type.py
+++ b/tests/unit/test_summarizer_context_type.py
@@ -44,7 +44,7 @@ class TestSummarizerContextType:
def test_session_uri_defaults_to_resource(self):
"""Session URIs should not be treated as memory-specific content."""
- uri = "viking://session/default/sess_001/history/archive_001"
+ uri = "viking://user/default/sessions/sess_001/history/archive_001"
assert context_type_for_uri(uri) == "resource"
def test_resource_named_memories_does_not_classify_as_memory(self):
diff --git a/tests/unit/test_uri_short_format.py b/tests/unit/test_uri_short_format.py
index 6e85e4f97..72c35dcbc 100644
--- a/tests/unit/test_uri_short_format.py
+++ b/tests/unit/test_uri_short_format.py
@@ -90,7 +90,6 @@ class TestVikingURIExplicitFormat:
[
("viking://resources", "resources"),
("viking://user", "user"),
- ("viking://session/abc123", "session"),
("viking://queue", "queue"),
("viking://temp", "temp"),
],
diff --git a/tests/unit/test_uri_validation.py b/tests/unit/test_uri_validation.py
index cabd1b94e..15d8fa773 100644
--- a/tests/unit/test_uri_validation.py
+++ b/tests/unit/test_uri_validation.py
@@ -13,7 +13,6 @@ from openviking_cli.exceptions import InvalidURIError
"uri",
[
"viking://resources/docs",
- "viking://session/s1",
"viking://agent/skills/demo/SKILL.md",
"viking://system/memory",
"viking://",
diff --git a/tests/unit/test_vectorize_file_strategy.py b/tests/unit/test_vectorize_file_strategy.py
index 036e3a98c..76187c887 100644
--- a/tests/unit/test_vectorize_file_strategy.py
+++ b/tests/unit/test_vectorize_file_strategy.py
@@ -1041,7 +1041,7 @@ async def test_index_resource_skips_session_namespace(monkeypatch):
),
)
await embedding_utils.index_resource(
- uri="viking://session/default/sess_001/history/archive_001",
+ uri="viking://user/default/sessions/sess_001/history/archive_001",
ctx=DummyReq(),
)
diff --git a/web-studio/src/routes/filesystem/-components/terminal-panel.tsx b/web-studio/src/routes/filesystem/-components/terminal-panel.tsx
index dc03dede3..5bafb870a 100644
--- a/web-studio/src/routes/filesystem/-components/terminal-panel.tsx
+++ b/web-studio/src/routes/filesystem/-components/terminal-panel.tsx
@@ -468,7 +468,7 @@ function sessionToRef(session: Pick<SessionListItem, 'session_id' | 'uri'>) {
return {
label: session.session_id,
meta: 'session',
- uri: session.uri || `viking://session/${session.session_id}`,
+ uri: session.uri || `viking://~/sessions/${session.session_id}`,
}
}
@@ -1173,7 +1173,7 @@ export function TerminalPanel({
id,
formatJson(result),
]),
- refs: [{ uri: `viking://session/${id}` }],
+ refs: [{ uri: `viking://~/sessions/${id}` }],
})
}
return
@@ -1531,7 +1531,7 @@ export function TerminalPanel({
const sessionRef = (id: string): ResourceRef => ({
label: id,
meta: 'session',
- uri: `viking://session/${id}`,
+ uri: `viking://~/sessions/${id}`,
})
switch (subcommand) {
@@ -1642,7 +1642,7 @@ export function TerminalPanel({
{
label: archiveId,
meta: 'archive',
- uri: `viking://session/${id}/history/${archiveId}`,
+ uri: `viking://~/sessions/${id}/history/${archiveId}`,
},
],
title: `/session archive ${id} ${archiveId}`,
@@ -1732,7 +1732,7 @@ export function TerminalPanel({
{
label: toolResultId,
meta: 'tool result',
- uri: `viking://session/${id}/tool-results/${toolResultId}`,
+ uri: `viking://~/sessions/${id}/tool-results/${toolResultId}`,
},
],
title: `/session tool-result ${id} ${toolResultId}`,
@@ -1765,7 +1765,7 @@ export function TerminalPanel({
{
label: toolResultId,
meta: 'tool search',
- uri: `viking://session/${id}/tool-results/${toolResultId}`,
+ uri: `viking://~/sessions/${id}/tool-results/${toolResultId}`,
},
],
title: `/session tool-search ${id} ${toolResultId}`,
diff --git a/web-studio/src/routes/resources/-lib/api.test.ts b/web-studio/src/routes/resources/-lib/api.test.ts
index fff8379cb..d1756e7fa 100644
--- a/web-studio/src/routes/resources/-lib/api.test.ts
+++ b/web-studio/src/routes/resources/-lib/api.test.ts
@@ -63,7 +63,7 @@ describe('fetchDirectorySidecarContent', () => {
describe('fetchFsList', () => {
it('requests newest entries before the server applies node_limit', async () => {
- await fetchFsList('viking://session', { nodeLimit: 200 })
+ await fetchFsList('viking://~/sessions', { nodeLimit: 200 })
expect(getFsLsMock).toHaveBeenCalledWith({
query: expect.objectContaining({
diff --git a/web-studio/src/routes/sessions/-components/memory-impact.test.tsx b/web-studio/src/routes/sessions/-components/memory-impact.test.tsx
index 5be22d464..98f144723 100644
--- a/web-studio/src/routes/sessions/-components/memory-impact.test.tsx
+++ b/web-studio/src/routes/sessions/-components/memory-impact.test.tsx
@@ -27,7 +27,7 @@ afterEach(cleanup)
const diffs: SessionMemoryDiff[] = [
{
archiveId: 'archive-1',
- archiveUri: 'viking://session/archive-1',
+ archiveUri: 'viking://user/default/sessions/archive-1',
extractedAt: '2026-07-26T00:00:00Z',
operations: [
{
@@ -47,7 +47,7 @@ const diffs: SessionMemoryDiff[] = [
},
{
archiveId: 'archive-2',
- archiveUri: 'viking://session/archive-2',
+ archiveUri: 'viking://user/default/sessions/archive-2',
extractedAt: '2026-07-26T01:00:00Z',
operations: [
{