feat: workspace.get_llm() and get_secrets() for OpenHandsCloudWorkspace credential inheritance by xingyaoww · Pull Request #2409 · OpenHands/software-agent-sdk

xingyaoww · 2026-03-13T13:01:02Z

Summary

Adds get_llm() and get_secrets() methods to OpenHandsCloudWorkspace, enabling SDK-created conversations to inherit the user's SaaS credentials.

Design

get_llm(**kwargs): Calls GET /api/v1/users/me?expose_secrets=true with both Bearer token and X-Session-API-Key headers (dual auth). Extracts llm_model, llm_api_key, llm_base_url and returns a fully usable LLM instance. User-provided kwargs override SaaS settings.
get_secrets(names=None): Calls GET /sandboxes/{id}/settings/secrets (X-Session-API-Key auth) for names only, then builds LookupSecret references with env_headers. Raw secret values never transit through the SDK client — they are resolved lazily by the agent-server inside the sandbox.

Security

get_llm() sends both the Bearer token (user identity) and the sandbox session key (active sandbox proof). The server verifies the session key belongs to a sandbox owned by the same user. This prevents a leaked API key from being used to extract raw LLM credentials.

Usage

with OpenHandsCloudWorkspace(...) as workspace:
    llm = workspace.get_llm()  # calls /users/me?expose_secrets=true
    agent = Agent(llm=llm, tools=get_default_tools())

    conversation = Conversation(agent=agent, workspace=workspace)
    conversation.update_secrets(workspace.get_secrets())  # LookupSecret refs
    conversation.send_message("Analyze this repo")

Companion PR

Server-side: feat: expose_secrets param on /users/me + sandbox-scoped secrets API OpenHands#13383 (expose_secrets param on /users/me + secrets endpoints)

Tests

8 workspace tests (get_llm, get_secrets: happy path, overrides, no-key, filtering, errors)

Resolves OpenHands/OpenHands#13268

Agent Server images for this PR

• GHCR package: https://github.com/OpenHands/agent-sdk/pkgs/container/agent-server

Variants & Base Images

Variant	Architectures	Base Image	Docs / Tags
java	amd64, arm64	`eclipse-temurin:17-jdk`	Link
python	amd64, arm64	`nikolaik/python-nodejs:python3.13-nodejs22`	Link
golang	amd64, arm64	`golang:1.21-bookworm`	Link

Pull (multi-arch manifest)

# Each variant is a multi-arch manifest supporting both amd64 and arm64
docker pull ghcr.io/openhands/agent-server:8fe0b03-python

Run

docker run -it --rm \
  -p 8000:8000 \
  --name agent-server-8fe0b03-python \
  ghcr.io/openhands/agent-server:8fe0b03-python

All tags pushed for this build

ghcr.io/openhands/agent-server:8fe0b03-golang-amd64
ghcr.io/openhands/agent-server:8fe0b03-golang_tag_1.21-bookworm-amd64
ghcr.io/openhands/agent-server:8fe0b03-golang-arm64
ghcr.io/openhands/agent-server:8fe0b03-golang_tag_1.21-bookworm-arm64
ghcr.io/openhands/agent-server:8fe0b03-java-amd64
ghcr.io/openhands/agent-server:8fe0b03-eclipse-temurin_tag_17-jdk-amd64
ghcr.io/openhands/agent-server:8fe0b03-java-arm64
ghcr.io/openhands/agent-server:8fe0b03-eclipse-temurin_tag_17-jdk-arm64
ghcr.io/openhands/agent-server:8fe0b03-python-amd64
ghcr.io/openhands/agent-server:8fe0b03-nikolaik_s_python-nodejs_tag_python3.13-nodejs22-amd64
ghcr.io/openhands/agent-server:8fe0b03-python-arm64
ghcr.io/openhands/agent-server:8fe0b03-nikolaik_s_python-nodejs_tag_python3.13-nodejs22-arm64
ghcr.io/openhands/agent-server:8fe0b03-golang
ghcr.io/openhands/agent-server:8fe0b03-java
ghcr.io/openhands/agent-server:8fe0b03-python

About Multi-Architecture Support

Each variant tag (e.g., 8fe0b03-python) is a multi-arch manifest supporting both amd64 and arm64
Docker automatically pulls the correct architecture for your platform
Individual architecture tags (e.g., 8fe0b03-python-amd64) are also available if needed

Add methods to OpenHandsCloudWorkspace that call the new SaaS API endpoints to retrieve the user's LLM configuration and custom secrets: - get_llm(**llm_kwargs): Fetches LLM settings from the user's SaaS account and returns a configured LLM instance. User kwargs override SaaS defaults. - get_secrets(names=None): Fetches custom secrets and returns a dict[str, str] compatible with conversation.update_secrets(). These methods enable SDK users to inherit their SaaS credentials while retaining full control over agent customization. Depends on OpenHands/OpenHands#13306 for the server-side API endpoints. Related: OpenHands/OpenHands#13268 Co-authored-by: openhands <openhands@all-hands.dev>

github-actions · 2026-03-13T13:01:31Z

Python API breakage checks — ✅ PASSED

Result: ✅ PASSED

Action log

github-actions · 2026-03-13T13:01:41Z

REST API breakage checks (OpenAPI) — ✅ PASSED

Result: ✅ PASSED

Action log

…t-backed configs SDK LLM changes: - LLM.api_key now accepts str | SecretStr | SecretSource | None - Validator passes through SecretSource instances; deserialises dicts - Serializer delegates to SecretSource.model_dump() for round-tripping - _get_litellm_api_key_value() resolves SecretSource.get_value() lazily - _init_model_info_and_caps() skips network for SecretSource api_key OpenHandsCloudWorkspace changes: - get_llm() calls sandbox-scoped /settings/llm (SESSION_API_KEY auth) and returns LLM with api_key=LookupSecret — raw key never reaches client - get_secrets() calls /settings/secrets for names, returns dict of LookupSecret instances pointing to per-secret endpoints - Added _send_settings_request() for SESSION_API_KEY-authenticated calls Co-authored-by: openhands <openhands@all-hands.dev>

github-actions · 2026-03-13T13:56:04Z

Coverage Report •

File	Stmts	Miss	Cover	Missing
openhands-sdk/openhands/sdk/conversation/impl
remote_conversation.py	616	105	82%	77, 79, 150, 177, 190, 192–195, 205, 227–228, 233–236, 319, 329–331, 337, 378, 520–523, 525, 545–549, 554–557, 560, 572–576, 732–733, 737–738, 752, 776–777, 796, 807–808, 828–831, 833–834, 858–860, 863–867, 869–870, 874, 876–884, 886, 923, 1053, 1121–1122, 1126, 1131–1135, 1141–1147, 1160–1161, 1196, 1252, 1259, 1265–1266, 1344–1345
openhands-sdk/openhands/sdk/secret
secrets.py	64	2	96%	85, 88
openhands-workspace/openhands/workspace/cloud
workspace.py	232	178	23%	32–34, 108–110, 116, 119–120, 128, 132, 134–139, 147–148, 150, 153, 156–158, 162, 165–166, 169, 172–173, 177, 180–182, 185, 191, 193–195, 204–207, 217, 220, 225, 227–228, 230–232, 234, 236–237, 240–243, 245–246, 248–249, 251, 253–255, 258–259, 263–274, 278–279, 281–282, 290–291, 293–295, 297, 308, 321–322, 324–327, 331, 334–335, 338–340, 342–350, 352, 356–357, 359–362, 364–365, 371–373, 375–382, 391, 396, 429, 431–432, 434, 440, 442–448, 451, 453, 486, 488–489, 491–492, 494–499, 505, 520–521, 523–525, 527–535, 537, 540, 543, 546
TOTAL	20300	5933	70%

LookupSecret now supports env_headers — a mapping of header name to environment variable name. Headers are resolved from os.environ at get_value() call time. This ensures the SESSION_API_KEY is never embedded in the serialized LookupSecret; only the env var *name* travels over the wire. Resolution only succeeds inside the sandbox where the env var is set. - LookupSecret: add env_headers field, merge into headers in get_value() - LLM: add assert to narrow dict type for pyright - Workspace: use env_headers instead of raw headers in get_llm/get_secrets - Tests: 3 new env_headers enforcement tests (28 total SDK tests) - Fix pyright errors in examples and tests Co-authored-by: openhands <openhands@all-hands.dev>

Per feedback, LookupSecret is only needed for secrets — not LLM config. Reverted all LLM.api_key SecretSource changes (api_key stays str|SecretStr|None). workspace.get_llm() now returns a real LLM with the raw api_key. Deleted test_llm_secret_source_api_key.py (no longer applicable). - LLM class: reverted to main (no SecretSource support) - workspace.get_llm(): builds LLM from plain JSON response - workspace.get_secrets(): still uses LookupSecret+env_headers - 8 workspace tests pass Co-authored-by: openhands <openhands@all-hands.dev>

…ings/llm The OpenHands app server already has GET /api/v1/users/me which returns full user settings including llm_model, llm_api_key, llm_base_url. A new expose_secrets query param returns the raw api_key. get_llm() now calls /users/me?expose_secrets=true via _send_api_request (Bearer token auth) instead of /settings/llm via _send_settings_request (X-Session-API-Key auth), eliminating the need for a dedicated LLM settings endpoint. Co-authored-by: openhands <openhands@all-hands.dev>

The server now requires a valid session key when expose_secrets=true. get_llm() includes the sandbox session key header alongside the Bearer token, proving there is an active sandbox owned by the caller. Co-authored-by: openhands <openhands@all-hands.dev>

Demonstrates workspace.get_llm() and workspace.get_secrets() APIs that allow SDK users to inherit LLM config and secrets from their OpenHands Cloud account without providing LLM_API_KEY separately. Co-authored-by: openhands <openhands@all-hands.dev>

…tifacts The cleanup() method was sending DELETE to the collection route (/api/v1/sandboxes?sandbox_id=X) which returned 405. The server expects DELETE /api/v1/sandboxes/{id} with sandbox_id as a required query parameter (FastAPI routing quirk). Fixed to use both. Also adds .pr/ folder with integration test logs and report from running 09_cloud_workspace_saas_credentials.py against the staging deployment of OpenHands PR #13383. Co-authored-by: openhands <openhands@all-hands.dev>

github-actions · 2026-03-16T17:59:02Z

✅ PR Artifacts Cleaned Up

The .pr/ directory has been automatically removed.

Two bugs fixed: 1. update_secrets() crashed with 'LookupSecret is not JSON serializable' when passing SecretSource objects from get_secrets(). Now uses model_dump(mode='json') for Pydantic SecretSource instances. 2. cleanup() was calling DELETE /api/v1/sandboxes?sandbox_id=X (405). Fixed to DELETE /api/v1/sandboxes/{id}?sandbox_id={id} matching the server's FastAPI route definition. Updated .pr/ logs from clean end-to-end run with 2 secrets (DUMMY_1, DUMMY_2) successfully discovered, injected, and used. Co-authored-by: openhands <openhands@all-hands.dev>

The old prompt only listed env var names — it didn't prove values actually resolved. New prompt asks the agent to print the last 50% of each secret's value, which revealed that LookupSecret refs are recognized by the agent-server (appear in system prompt) but values are not yet injected as real environment variables. Co-authored-by: openhands <openhands@all-hands.dev>

- LookupSecret HTTP resolution works from both SDK client and inside sandbox - Identified deployment dependency: env_headers requires updated agent-server image - Identified SecretStr serialization redaction issue with update_secrets() - Proved _export_envs pipeline works end-to-end with plain string secrets Co-authored-by: openhands <openhands@all-hands.dev>

Remove the env_headers field from LookupSecret. Instead, pass the session key directly in the existing headers field and use expose_secrets=True context in model_dump() so SecretStr header values survive JSON serialization through update_secrets(). This eliminates a deployment dependency — the existing agent-server image (1.13.0-python) already supports LookupSecret.headers, so no image update is needed. Changes: - secrets.py: Remove env_headers field and os.environ resolution - workspace.py: Use headers={} with actual session key, remove _env_headers() - remote_conversation.py: Add context={'expose_secrets': True} to model_dump() - test: Update assertion from env_headers to headers Co-authored-by: openhands <openhands@all-hands.dev>

Co-authored-by: openhands <openhands@all-hands.dev>

Agent output from inside sandbox: DUMMY_1: len=14, last_half=ecret 1 DUMMY_2: len=14, last_half=ecret 2 Co-authored-by: openhands <openhands@all-hands.dev>

Fresh run_combined.log from end-to-end test showing: DUMMY_1: len=14, last_half=ecret 1 DUMMY_2: len=14, last_half=ecret 2 Co-authored-by: openhands <openhands@all-hands.dev>

The staging server requires PR #13383 for the /sandboxes/{id}/settings/secrets endpoints. Only the sandbox agent-server is stock 1.13.0-python. Co-authored-by: openhands <openhands@all-hands.dev>

xingyaoww · 2026-03-16T20:56:05Z

@OpenHands please fix the failing CIs (pre-commits, duplicated example number) and open a new PR on OpenHands/docs repo for the docs example. make sure you also modify the docs related to OpenHandsCloudWorkspace there to add the ability of get llm & secrets.

openhands-ai · 2026-03-16T20:56:24Z

I'm on it! xingyaoww can track my progress at all-hands.dev

all-hands-bot

🟢 Good taste - Elegant credential inheritance design. The lazy-resolved LookupSecret pattern keeps security boundaries clean (raw values never transit through SDK). Code is simple, well-tested, and solves a real problem.

openhands-sdk/openhands/sdk/secret/secrets.py

The method always returns response.text (which is always str) or raises via raise_for_status(), so str | None was imprecise. Co-authored-by: openhands <openhands@all-hands.dev>

…long - Rename 09_cloud_workspace_saas_credentials.py to 10_ to avoid duplicate number conflict with 09_acp_agent_with_remote_runtime.py - Fix E501 in test docstring (shorten description) - Fix E501 in example comments and string literals (wrap lines) Co-authored-by: openhands <openhands@all-hands.dev>

- Add 'Inheriting SaaS Credentials' section documenting get_llm() and get_secrets() methods for credential inheritance - Add SaaS Credentials example section with the full expandable example - Update page description to mention SaaS credential inheritance Companion SDK PR: OpenHands/software-agent-sdk#2409 Co-authored-by: openhands <openhands@all-hands.dev>

openhands-ai · 2026-03-16T21:06:23Z

Based on my review of the work done, here's the summary:

Summary of Changes

SDK PR Fixes (pushed to `feat/cloud-workspace-get-llm-secrets` on OpenHands/software-agent-sdk)

All changes have been pushed. The commit 10cbd942 addressed the three failing CI checks:

✅ Pre-commit (E501 line too long) — Fixed 3 lines:
- Shortened a test docstring in tests/workspace/test_cloud_workspace_sdk_settings.py
- Wrapped a long comment and a long string literal in the example file
✅ Duplicate example number — Renamed 09_cloud_workspace_saas_credentials.py → 10_cloud_workspace_saas_credentials.py (09 was already taken by 09_acp_agent_with_remote_runtime.py on main)
✅ Undocumented example — Resolved by creating the docs PR below

All three previously-failing CI checks (pre-commit, check-duplicates, check-examples) are now passing.

Docs PR (OpenHands/docs#401)

Created and pushed: OpenHands/docs#401

Updated sdk/guides/agent-server/cloud-workspace.mdx with:

"Inheriting SaaS Credentials" section documenting get_llm() and get_secrets() with code snippets
"SaaS Credentials Example" section with the full expandable example code block (referencing examples/02_remote_agent_server/10_cloud_workspace_saas_credentials.py)
Updated page description metadata

Also pushed a matching branch name (feat/cloud-workspace-get-llm-secrets) to the docs repo so the SDK's check-examples CI can discover the documentation.

Checklist

Fix pre-commit failures (E501 line too long)
Fix duplicated example number
Open new PR on OpenHands/docs repo for the example documentation
Modify cloud-workspace docs to add get_llm() & get_secrets() capability documentation
All changes pushed to remote branches

openhands-workspace/openhands/workspace/cloud/workspace.py

Co-authored-by: openhands <openhands@all-hands.dev>

xingyaoww · 2026-03-17T11:41:05Z

✅ E2E Testing Complete (2026-03-17)

Ran full e2e against the staging feature deployment (ohpr-13383-240.staging.all-hands.dev) with the latest server commit c0ed13f from OpenHands/OpenHands#13383.

Results — all passing

Phase	Status	Details
`workspace.get_llm()`	✅	`model=litellm_proxy/minimax-m2.5`, api_key + base_url inherited from SaaS
`workspace.get_secrets()`	✅	Returns 3 secrets: `['secret_1', 'secret_2', 'github_token']`
Provider token discovery	✅	`github_token` included — provider tokens now flow to SDK
`LookupSecret` resolution	✅	All 3 secrets resolved: `secret_1` (len=5), `secret_2` (len=5), `github_token` (len=40)
Sandbox cleanup	✅	Sandbox created + deleted cleanly

Test artifacts pushed to .pr/ folder in commit d5367f5:

.pr/README.md — full report
.pr/logs/e2e_provider_tokens.log — raw output

The server companion PR (#13383) was deployed via deploy#3436 and the get_provider_tokens(as_env_vars=True) refactor works end-to-end.

Co-authored-by: openhands <openhands@all-hands.dev>

xingyaoww · 2026-03-17T12:07:29Z

📋 Full Example 10 stdout uploaded

Ran examples/02_remote_agent_server/10_cloud_workspace_saas_credentials.py end-to-end against ohpr-13383-240.staging.all-hands.dev.

Complete stdout log pushed in commit d8b4ac4: .pr/logs/example_10_full_stdout.log

Highlights from the run:

Sandbox 5ePS0rOBeWHQci1yUZEqMv created → RUNNING in ~60s
LLM: litellm_proxy/minimax-m2.5 inherited from SaaS
Secrets discovered: ['secret_1', 'secret_2', 'github_token']
3 secrets injected into conversation via update_secrets()
Agent resolved all 3 env vars inside sandbox and wrote SECRETS_CHECK.txt
Run completed in 34.2s, 17 events, sandbox cleaned up

examples/02_remote_agent_server/10_cloud_workspace_share_credentials.py

…e_credentials Co-authored-by: openhands <openhands@all-hands.dev>

Matches rename in OpenHands/software-agent-sdk#2409. Co-authored-by: openhands <openhands@all-hands.dev>

xingyaoww · 2026-03-17T13:06:49Z

Besides agent confirmation, i've also confirmed that this is working by running the script manually.
Good to merge after one final confirmation! (cc @malhotra5)

openhands-workspace/openhands/workspace/cloud/workspace.py

malhotra5

LGTM! Nice work, just a nit regarding error handling on the API calls

Add tenacity retry decorators to get_llm() and _send_settings_request() with up to 3 attempts and exponential backoff. Only transient errors (network issues, server 5xx) are retried; client errors (4xx) fail immediately. Co-authored-by: openhands <openhands@all-hands.dev>

This was referenced Mar 13, 2026

feat: expose_secrets param on /users/me + sandbox-scoped secrets API OpenHands/OpenHands#13383

Merged

feat: SDK-created conversations should inherit SaaS settings (credentials, repo context) OpenHands/OpenHands#13268

Closed

xingyaoww changed the title ~~DRAFT: feat: add get_llm() and get_secrets() to OpenHandsCloudWorkspace~~ feat: LLM api_key accepts SecretSource; workspace returns LookupSecret-backed configs Mar 13, 2026

openhands-agent added 2 commits March 13, 2026 14:14

xingyaoww changed the title ~~feat: LLM api_key accepts SecretSource; workspace returns LookupSecret-backed configs~~ feat: workspace.get_llm() and get_secrets() for SaaS credential inheritance Mar 13, 2026

openhands-agent added 4 commits March 13, 2026 16:04

openhands-agent added 5 commits March 16, 2026 18:13

docs: finalize test artifact with all results

4972ae9

Co-authored-by: openhands <openhands@all-hands.dev>

xingyaoww changed the title ~~feat: workspace.get_llm() and get_secrets() for SaaS credential inheritance~~ feat: workspace.get_llm() and get_secrets() for OpenHandsCloudWorkspace credential inheritance Mar 16, 2026

openhands-agent added 3 commits March 16, 2026 20:51

docs: update test artifact with unambiguous verification output

3deba1e

Agent output from inside sandbox: DUMMY_1: len=14, last_half=ecret 1 DUMMY_2: len=14, last_half=ecret 2 Co-authored-by: openhands <openhands@all-hands.dev>

docs: replace stale .pr/logs with final passing test output

6552831

Fresh run_combined.log from end-to-end test showing: DUMMY_1: len=14, last_half=ecret 1 DUMMY_2: len=14, last_half=ecret 2 Co-authored-by: openhands <openhands@all-hands.dev>

fix: clarify SaaS server is custom build from PR #13383

6f3c8de

The staging server requires PR #13383 for the /sandboxes/{id}/settings/secrets endpoints. Only the sandbox agent-server is stock 1.13.0-python. Co-authored-by: openhands <openhands@all-hands.dev>

xingyaoww marked this pull request as ready for review March 16, 2026 20:56

all-hands-bot approved these changes Mar 16, 2026

View reviewed changes

openhands-sdk/openhands/sdk/secret/secrets.py Outdated Show resolved Hide resolved

chore: Remove PR-only artifacts [automated]

84770cd

openhands-agent added 2 commits March 16, 2026 21:00

refactor: narrow LookupSecret.get_value() return type to str

7fd6d48

The method always returns response.text (which is always str) or raises via raise_for_status(), so str | None was imprecise. Co-authored-by: openhands <openhands@all-hands.dev>

xingyaoww mentioned this pull request Mar 16, 2026

docs: add get_llm() and get_secrets() to OpenHands Cloud Workspace guide OpenHands/docs#401

Open

malhotra5 reviewed Mar 16, 2026

View reviewed changes

openhands-workspace/openhands/workspace/cloud/workspace.py Show resolved Hide resolved

docs: add provider tokens e2e test results to .pr/

d5367f5

Co-authored-by: openhands <openhands@all-hands.dev>

docs: add full stdout of example 10 (SaaS credentials e2e)

d8b4ac4

Co-authored-by: openhands <openhands@all-hands.dev>

xingyaoww commented Mar 17, 2026

View reviewed changes

examples/02_remote_agent_server/10_cloud_workspace_share_credentials.py Show resolved Hide resolved

rename: 10_cloud_workspace_saas_credentials → 10_cloud_workspace_shar…

8755c4d

…e_credentials Co-authored-by: openhands <openhands@all-hands.dev>

xingyaoww pushed a commit to OpenHands/docs that referenced this pull request Mar 17, 2026

docs: update example filename saas_credentials → share_credentials

bb4ec36

Matches rename in OpenHands/software-agent-sdk#2409. Co-authored-by: openhands <openhands@all-hands.dev>

This was referenced Mar 17, 2026

docs: add cross-repo testing skill for SDK ↔ OH Cloud e2e workflow #2477

Open

docs: add cross-repo testing skill for SDK ↔ OH Cloud e2e workflow OpenHands/OpenHands#13446

Open

xingyaoww added 2 commits March 17, 2026 20:59

We should not override LLM / simplify prompt

066dde6

Simplify prompt

f9e0cd3

Merge branch 'main' into feat/cloud-workspace-get-llm-secrets

15104c4

malhotra5 reviewed Mar 17, 2026

View reviewed changes

openhands-workspace/openhands/workspace/cloud/workspace.py Show resolved Hide resolved

malhotra5 approved these changes Mar 17, 2026

View reviewed changes

allhands-bot and others added 2 commits March 17, 2026 14:48

chore: Remove PR-only artifacts [automated]

e69170a

Conversation

xingyaoww commented Mar 13, 2026 • edited by github-actions bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Design

Security

Usage

Companion PR

Tests

Uh oh!

github-actions bot commented Mar 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Python API breakage checks — ✅ PASSED

Uh oh!

github-actions bot commented Mar 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

REST API breakage checks (OpenAPI) — ✅ PASSED

Uh oh!

github-actions bot commented Mar 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Mar 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

xingyaoww commented Mar 16, 2026

Uh oh!

openhands-ai bot commented Mar 16, 2026

Uh oh!

all-hands-bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

openhands-ai bot commented Mar 16, 2026

Summary of Changes

SDK PR Fixes (pushed to feat/cloud-workspace-get-llm-secrets on OpenHands/software-agent-sdk)

Docs PR (OpenHands/docs#401)

Checklist

Uh oh!

Uh oh!

xingyaoww commented Mar 17, 2026

✅ E2E Testing Complete (2026-03-17)

Results — all passing

Uh oh!

xingyaoww commented Mar 17, 2026

📋 Full Example 10 stdout uploaded

Uh oh!

Uh oh!

xingyaoww commented Mar 17, 2026

Uh oh!

Uh oh!

malhotra5 left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

xingyaoww commented Mar 13, 2026 •

edited by github-actions bot

Loading

github-actions bot commented Mar 13, 2026 •

edited

Loading

github-actions bot commented Mar 13, 2026 •

edited

Loading

github-actions bot commented Mar 13, 2026 •

edited

Loading

github-actions bot commented Mar 16, 2026 •

edited

Loading

SDK PR Fixes (pushed to `feat/cloud-workspace-get-llm-secrets` on OpenHands/software-agent-sdk)