Skip to content

Commit 9f67577

Browse files
authored
docs: updates for #1250 (#1251)
1 parent 4f89a6d commit 9f67577

3 files changed

Lines changed: 9 additions & 9 deletions

File tree

README.md

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -48,20 +48,20 @@ npm install @openai/agents zod
4848

4949
```js
5050
import { run } from '@openai/agents';
51-
import { gitRepo, Manifest, SandboxAgent } from '@openai/agents/sandbox';
51+
import { gitRepo, SandboxAgent } from '@openai/agents/sandbox';
5252
import { UnixLocalSandboxClient } from '@openai/agents/sandbox/local';
5353

5454
const agent = new SandboxAgent({
5555
name: 'Workspace Assistant',
5656
instructions: 'Inspect the sandbox workspace before answering.',
57-
defaultManifest: new Manifest({
57+
defaultManifest: {
5858
entries: {
5959
repo: gitRepo({
6060
repo: 'openai/openai-agents-js',
6161
ref: 'main',
6262
}),
6363
},
64-
}),
64+
},
6565
});
6666

6767
const result = await run(

docs/src/content/docs/guides/sandbox-agents/concepts.mdx

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -117,7 +117,7 @@ The main SDK pieces map onto those layers like this:
117117

118118
A practical design order is:
119119

120-
1. Define the fresh-session workspace contract with `Manifest`.
120+
1. Define the fresh-session workspace contract with a `Manifest` or manifest init object.
121121
2. Define the agent with `SandboxAgent`.
122122
3. Add built-in or custom capabilities.
123123
4. Decide how each run should obtain its sandbox session in `run(agent, input, { sandbox: ... })` or `new Runner({ sandbox: ... })`.
@@ -150,7 +150,7 @@ Sandbox client choice, sandbox-session reuse, manifest override, and snapshot se
150150

151151
### `defaultManifest`
152152

153-
`defaultManifest` is the default `Manifest` used when the runner creates a fresh sandbox session for this agent. Use it for the files, repos, helper material, output directories, and mounts the agent should usually start with.
153+
`defaultManifest` is the default workspace used when the runner creates a fresh sandbox session for this agent. Pass either a `Manifest` instance or the same init object you would pass to `new Manifest(...)`. Use it for the files, repos, helper material, output directories, and mounts the agent should usually start with.
154154

155155
This is only the default. A run can override it with `sandbox.manifest`, and a reused or resumed sandbox session keeps its existing workspace state.
156156

@@ -332,7 +332,7 @@ These options only matter when the runner is creating a fresh sandbox session:
332332

333333
| Option | Use it when | Notes |
334334
| --- | --- | --- |
335-
| `manifest` | You want a one-off fresh-session workspace override. | Falls back to `agent.defaultManifest` when omitted. |
335+
| `manifest` | You want a one-off fresh-session workspace override. | Accepts a `Manifest` or manifest init object. Falls back to `agent.defaultManifest` when omitted. |
336336
| `snapshot` | A fresh sandbox session should be seeded from a snapshot. | Useful for resume-like flows or remote snapshot clients. |
337337
| `options` | The sandbox client needs creation-time options. | Common for Docker images, provider timeouts, and similar client-specific settings. |
338338

examples/docs/toppage/sandboxAgent.ts

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1,14 +1,14 @@
11
import { run } from '@openai/agents';
2-
import { gitRepo, Manifest, SandboxAgent } from '@openai/agents/sandbox';
2+
import { gitRepo, SandboxAgent } from '@openai/agents/sandbox';
33
import { UnixLocalSandboxClient } from '@openai/agents/sandbox/local';
44

55
const agent = new SandboxAgent({
66
name: 'Workspace Assistant',
77
model: 'gpt-5.5',
88
instructions: 'Inspect the repo before changing files.',
9-
defaultManifest: new Manifest({
9+
defaultManifest: {
1010
entries: { repo: gitRepo({ repo: 'openai/openai-agents-js' }) },
11-
}),
11+
},
1212
});
1313

1414
const result = await run(

0 commit comments

Comments
 (0)