fix(openhuman): restore AgentTeams runtime contract

[RerankerGuo]

Summary

• make the OpenHuman entrypoint consume the terminal AGENTTEAMS_* worker, storage, Matrix, gateway, and controller contract
• read the renamed agentteams-gateway provider from generated openclaw.json
• restore OpenHuman's dedicated Docker/Kubernetes working directory and Kubernetes runtime label
• point Helm at the published agentteams-openhuman-worker image

Root cause

The backend normalization in #998 removed the existing OpenHuman working-directory and runtime-label cases while adding QwenPaw. The subsequent AgentTeams hard rename updated the controller contract but did not update the OpenHuman entrypoint or its Helm image default. A current controller therefore injects AGENTTEAMS_WORKER_NAME, while the entrypoint exits immediately requiring HICLAW_WORKER_NAME; it also looks for the retired storage alias and gateway provider.

Impact

OpenHuman workers can start from current Docker and Kubernetes controller requests, pull their workspace from the AgentTeams storage prefix, build Matrix/LLM configuration from current fields, and report readiness with current controller auth settings. Other worker runtimes are unchanged.

Validation

• go test ./... in hiclaw-controller/ with Go 1.25.0
• bash openhuman/tests/test-entrypoint-agentteams-env.sh
• bash -n openhuman/scripts/openhuman-worker-entrypoint.sh openhuman/tests/test-entrypoint-agentteams-env.sh
• Helm 3.14.0 dependency build, lint, and template dry run using the repository workflow values
• git diff --check

[github-actions]

❌ Integration Tests Failed (llm-interaction-2 / mgr=copaw / wk=hermes)

Commit: ff82011
Workflow run: #1659

Test Results

No test output captured.

Debug Log (tail)

No debug logs available.

---

📦 Download full debug logs & test artifacts

[github-actions]

❌ Integration Tests Failed (llm-interaction / mgr=openclaw / wk=openclaw)

Commit: ff82011
Workflow run: #1659

Test Results

No test output captured.

Debug Log (tail)

No debug logs available.

---

📦 Download full debug logs & test artifacts

[github-actions]

❌ Integration Tests Failed (controller-cr / mgr=openclaw / wk=openclaw)

Commit: ff82011
Workflow run: #1659

Test Results

No test output captured.

Debug Log (tail)

No debug logs available.

---

📦 Download full debug logs & test artifacts

[github-actions]

❌ Integration Tests Failed (controller-cr / mgr=copaw / wk=copaw)

Commit: ff82011
Workflow run: #1659

Test Results

No test output captured.

Debug Log (tail)

No debug logs available.

---

📦 Download full debug logs & test artifacts

[github-actions]

❌ Integration Tests Failed (llm-interaction-2 / mgr=openclaw / wk=openclaw)

Commit: ff82011
Workflow run: #1659

Test Results

No test output captured.

Debug Log (tail)

No debug logs available.

---

📦 Download full debug logs & test artifacts

[github-actions]

❌ Integration Tests Failed (llm-interaction / mgr=copaw / wk=hermes)

Commit: ff82011
Workflow run: #1659

Test Results

No test output captured.

Debug Log (tail)

No debug logs available.

---

📦 Download full debug logs & test artifacts

[github-actions]

❌ Integration Tests Failed (llm-interaction / mgr=copaw / wk=copaw)

Commit: ff82011
Workflow run: #1659

Test Results

No test output captured.

Debug Log (tail)

No debug logs available.

---

📦 Download full debug logs & test artifacts

[github-actions]

❌ Integration Tests Failed (controller-cr / mgr=copaw / wk=hermes)

Commit: ff82011
Workflow run: #1659

Test Results

No test output captured.

Debug Log (tail)

No debug logs available.

---

📦 Download full debug logs & test artifacts

[github-actions]

❌ Integration Tests Failed (llm-interaction-2 / mgr=copaw / wk=copaw)

Commit: ff82011
Workflow run: #1659

Test Results

No test output captured.

Debug Log (tail)

No debug logs available.

---

📦 Download full debug logs & test artifacts

[github-actions]

❌ Integration Tests Failed (controller-cr-2 / mgr=copaw / wk=copaw)

Commit: ff82011
Workflow run: #1659

Test Results

No test output captured.

Debug Log (tail)

No debug logs available.

---

📦 Download full debug logs & test artifacts

[RerankerGuo]

CI attribution: all 9 translation/base/image build checks passed. I inspected the raw logs for all 10 failed integration jobs in run #1659; every matrix entry exits in install-embedded before running its shard with the same base-workflow error:

[AgentTeams ERROR] LLM API Key is required (set via environment variable in non-interactive mode)

The branch-specific verification remains green locally (go test ./..., OpenHuman entrypoint regression, and Helm 3.14 lint/template). This CI failure is therefore the repository secret/env contract issue addressed by #1026, not an OpenHuman regression introduced by this PR.

[RerankerGuo]

Rebuilt on current main and removed changes already absorbed by the AgentTeams rename. The remaining diff is limited to current contract gaps: dedicated OpenHuman working directories in Docker/Kubernetes, runtime fallback coverage, Matrix user-id derivation, inline/projected auth-token handling, and a focused entrypoint environment test.

Verification:

• /opt/homebrew/opt/go/libexec/bin/go test ./internal/backend — pass
• /bin/bash openhuman/tests/test-entrypoint-agentteams-env.sh — pass
• /bin/bash -n openhuman/scripts/openhuman-worker-entrypoint.sh openhuman/tests/test-entrypoint-agentteams-env.sh
• git diff --check

The rebuilt branch does not restore old /opt/hiclaw, Helm, or Dockerfile changes.

[RerankerGuo]

Fixed the fresh helm-lint failure: the test itself contained a split retired-brand literal, which tests/check-agentteams-rename-defaults.sh correctly rejects. Removed that redundant assertion; repository-level rename validation remains authoritative.

Reverification:

• bash tests/check-agentteams-rename-defaults.sh — pass
• bash openhuman/tests/test-entrypoint-agentteams-env.sh — pass
• go test ./internal/backend — pass