docs(hermes): fix agent-specific guidance

[miyoungc]

Summary

Fixes two QA-reported Hermes documentation defects.
The Hermes provider-switch page now excludes OpenClaw gateway behavior, and the Network Policies page no longer repeats the Hermes baseline-policy guidance.

Related Issue

Fixes #7902.
Fixes #7903.

Changes

• Scope the OpenClaw and Hermes provider-switch behavior to their respective guide variants.
• Remove the duplicate Hermes baseline-policy callout.
• Add regression tests for both generated Hermes pages.

Type of Change

• Code change (feature, bug fix, or refactor)
• Code change with doc updates
• Doc only (prose changes, no code sample modifications)
• Doc only (includes code sample changes)

Quality Gates

• Tests added or updated for changed behavior
• Existing tests cover changed behavior — justification:
• Tests not applicable — justification:
• Docs updated for user-facing behavior changes
• Docs not applicable — justification:
• Sensitive paths changed (security, policy, credentials, preflight, onboarding, inference, runner, sandbox, or messaging)
• Sensitive-path review completed or maintainer-approved waiver recorded — reviewer/approval link/justification: Independent documentation writer review passed on d6eeb445e; the diff changes agent-specific inference and policy documentation only and does not change runtime behavior.
• Non-success, skipped, or missing CI check accepted by maintainer — check name, approval link, and follow-up issue:

Documentation Writer Review

• Documentation writer subagent reviewed the completed changes
• Result: docs-updated
• Evidence: Reviewed docs/inference/switch-providers.mdx and docs/reference/network-policies.mdx against the writing rules and documentation style. The focused regression test passed 17 of 17 tests, and npm run docs --silent completed with 0 errors.
• Agent: Codex Desktop

DGX Station Hardware Evidence

• Tested on DGX Station
• Tested commit:
• Station profile/scenario:
• Result:
• Supporting evidence:

Verification

• PR description includes a Signed-off-by: line and every commit appears as Verified in GitHub
• Normal pre-commit, commit-msg, and pre-push hooks passed, or npm run validate:pr passed after refreshing origin/main when hooks were skipped or unavailable
• Targeted behavior tests pass for the current change set, or tests are marked not applicable above — command/result or justification: npx vitest run test/agent-variant-docs.test.ts passed 17 of 17 tests.
• Applicable broad gate passed — npm test for broad runtime/test-harness changes; npm run check for repo-wide validation/coverage changes — command/result: Not run because the diff changes documentation and focused regression assertions only.
• Quality Gates section completed with required justifications or waivers
• No secrets, API keys, or credentials committed
• npm run docs builds without warnings (doc changes only) — Fern completed with 0 errors and reported one light-mode accent-color contrast warning in unchanged configuration.
• Doc pages follow the style guide (doc changes only)
• New doc pages include SPDX header and frontmatter (new pages only)

---

Signed-off-by: Miyoung Choi miyoungc@nvidia.com

Summary by CodeRabbit

• Documentation
  • Updated the “switch providers at runtime” workflow to show distinct, provider-specific instructions, with shared follow-up guidance applied after.
  • Refined the Hermes “Baseline Policy” section formatting by removing an extra wrapper while keeping the displayed text unchanged.
• Tests
  • Added assertions to ensure Hermes docs exclude OpenClaw gateway-behavior phrases and that Hermes-only baseline-policy guidance appears exactly once.

[coderabbitai]

📝 Walkthrough

Walkthrough

The documentation separates OpenClaw and Hermes provider-switch guidance, removes the Hermes baseline-policy wrapper causing repeated content, and adds tests for provider-specific output and single-occurrence policy text.

Changes

Agent variant documentation

Layer / File(s)	Summary
Provider switch guidance docs/inference/switch-providers.mdx, test/agent-variant-docs.test.ts	OpenClaw and Hermes runtime provider-switch instructions are separated into AgentOnly blocks, with tests verifying variant-specific content.
Hermes policy rendering docs/reference/network-policies.mdx, test/agent-variant-docs.test.ts	The Hermes baseline-policy text is no longer wrapped in Note, and a test verifies the explanation appears exactly once.

Estimated code review effort: 2 (Simple) | ~10 minutes

Possibly related PRs

• NVIDIA/NemoClaw#7892: Updates related Hermes dashboard-mirroring guidance in the same provider-switch documentation.

Suggested labels: integration: hermes, bug-fix

Suggested reviewers: cv, prekshivyas, senthilr-nv

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	The docs and regression tests address both linked issues: Hermes-only provider guidance and the duplicated baseline-policy paragraph.
Out of Scope Changes check	✅ Passed	The changes stay within the documented Hermes guidance and test coverage scope with no obvious unrelated edits.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly reflects the main change: fixing Hermes-specific documentation guidance.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch codex/fix-qa-hermes-docs

---

_{Comment @coderabbitai help to get the list of available commands.}

[github-code-quality]

Code Coverage Overview

Languages: TypeScript

TypeScript / code-coverage/plugin

The overall coverage in commit d6eeb44 in the codex/fix-qa-hermes-... branch remains at 96%, unchanged from commit 148efea in the main branch.

TypeScript / code-coverage/cli

The overall coverage in commit d6eeb44 in the codex/fix-qa-hermes-... branch remains at 81%, unchanged from commit 148efea in the main branch.

Show a code coverage summary of the most impacted files.

File	main 148efea	codex/fix-qa-hermes-... d6eeb44	+/-
src/lib/domain/.../connect-env.ts	97%	89%	-8%
src/lib/sandbox...rce-identity.ts	88%	88%	0%
src/lib/tunnel/services.ts	73%	73%	0%

---

_{Updated July 30, 2026 20:42 UTC}

[github-actions]

🌿 Preview your docs: https://nvidia-preview-pr-7941.docs.buildwithfern.com/nemoclaw

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@test/agent-variant-docs.test.ts`:
- Around line 295-310: Expand the test around renderAgentVariantPage to assert
that the Hermes output excludes the OpenClaw configuration and hot-reload
guidance markers from the source, in addition to the existing gateway sentence
assertion. Keep the OpenClaw positive assertion and Hermes-specific positive
assertion unchanged.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 115e8506-282f-4abc-a1d0-c87241702ee0

📥 Commits

Reviewing files that changed from the base of the PR and between 148efea and 308cc55.

📒 Files selected for processing (3)

• docs/inference/switch-providers.mdx
• docs/reference/network-policies.mdx
• test/agent-variant-docs.test.ts

💤 Files with no reviewable changes (1)

• docs/reference/network-policies.mdx

[github-actions]

PR Review Advisor — No blocking findings reported

Advisor assessment: No blocking advisor findings reported
Next action: No advisor follow-up needed.
Findings: 0 blockers · 0 warnings · 0 suggestions

Model lanes

• GPT-5.6 Terra (primary): Completed · high confidence · 0 blockers · 0 warnings · 0 suggestions
• Nemotron 3 Ultra (second opinion): Completed · high confidence · 0 blockers · 0 warnings · 0 suggestions
• Model comparison: normalized findings match; normalized E2E selections match; severity counts match.

Nemotron output stays in workflow artifacts and does not change the assessment above.

E2E guidance

Advisory only. E2E / PR Gate selects and runs jobs independently.

Recommended E2E: None

Workflow run details

This automated review informs maintainers. Warnings and suggestions do not require a response. A maintainer decides whether to merge.