docs(docs): fix links to files outside documentation routes (#1648)

# Pull Request

## Description

<!-- Brief description of changes. Link related issues using Closes #123
-->

Fix the Docusaurus build failure from [run
35765089888](https://github.com/microsoft/physical-ai-toolchain/actions/runs/35765089888)
by marking repository-external documentation links with the supported
`pathname://` protocol.

Six links across five documentation pages now bypass same-plugin
Markdown resolution while preserving their link text, target paths, and
fragments. Strict broken-link enforcement remains enabled; no Docusaurus
configuration, workflow, dependency, test, or generated file changes are
included.

Closes #1647 

## Type of Change
<!-- Mark relevant options with [x] -->

- [x] 🐛 Bug fix (non-breaking change fixing an issue)
- [x] ✨ New feature (non-breaking change adding functionality)
- [ ] 💥 Breaking change (fix or feature causing existing functionality
to change)
- [x] 📚 Documentation update
- [ ] 🏗️ Infrastructure change (Terraform/IaC)
- [ ] ♻️ Refactoring (no functional changes)

## Component(s) Affected
<!-- Mark all that apply -->

- [ ] `infrastructure/terraform/prerequisites/` - Azure subscription
setup
- [ ] `infrastructure/terraform/` - Terraform infrastructure
- [ ] `infrastructure/setup/` - OSMO control plane / Helm
- [x] `workflows/` - Training and evaluation workflows
- [ ] `training/` - Training pipelines and scripts
- [x] `docs/` - Documentation

## Testing Performed
<!-- Describe testing. Check applicable items -->

- [ ] Terraform `plan` reviewed (no unexpected changes)
- [ ] Terraform `apply` tested in dev environment
- [ ] Training scripts tested locally with Isaac Sim
- [ ] OSMO workflow submitted successfully
- [ ] Smoke tests passed (`smoke_test_azure.py`)

| Check | Result |
|-------|--------|
| Docusaurus unit tests | 5 suites and 20 tests passed |
| Docusaurus production build | Client and server compiled; static files
generated |
| Markdown lint | 227 files checked; 0 issues |
| Spell check | 901 files checked; 0 issues |
| `git diff --check origin/main...HEAD` | Passed |

The production build continues to report three pre-existing, non-fatal
anchor warnings outside this change.

## Documentation Impact
<!-- Select one -->

- [ ] No documentation changes needed
- [x] Documentation updated in this PR
- [ ] Documentation issue filed

## Bug Fix Checklist

*Complete this section for bug fix PRs. Skip for other contribution
types.*

- [ ] Linked to issue being fixed
- [ ] Regression test included, OR
- [x] Justification for no regression test: The existing Docusaurus
production build is the regression gate for static link resolution and
passed with strict broken-link checks enabled.

## Checklist

- [x] My code follows the [project conventions](copilot-instructions.md)
- [x] Commit messages follow [conventional commit
format](instructions/commit-message.instructions.md)
- [x] I have performed a self-review
- [x] Documentation impact assessed above
- [x] No new linting warnings introduced

Author: jkim323

docs/contributing/README.md

@@ -3,7 +3,7 @@ sidebar_position: 1
 title: Contributing to Physical AI Toolchain
 description: Guide for contributing including prerequisites, deployment validation, and style conventions
 author: Microsoft Robotics-AI Team
-ms.date: 2026-09-21
+ms.date: 2026-09-22
 ms.topic: how-to
 keywords:
   - azure
@@ -114,7 +114,7 @@ This project uses [Conventional Commits](https://www.conventionalcommits.org/):
 
 Types: `feat`, `fix`, `refactor`, `perf`, `style`, `test`, `docs`, `build`, `ops`, `chore`, `security`.
 
-Select a scope from the [commit-message instructions](../../.github/instructions/commit-message.instructions.md#scopes), such as `infrastructure`, `training`, `scripts`, `docs`, or `build`.
+Select a scope from the [commit-message instructions](pathname://../../.github/instructions/commit-message.instructions.md#scopes), such as `infrastructure`, `training`, `scripts`, `docs`, or `build`.
 
 Use imperative descriptions and keep the subject under 100 bytes. For larger changes, use at most five body bullets totaling fewer than 300 bytes. End with an emoji attribution footer followed by `- Generated by Copilot`.
 
@@ -189,7 +189,7 @@ npm run lint:sh
 npm run lint:md
 ```
 
-Run the applicable tests listed in [Testing Requirements](../../CONTRIBUTING.md#testing-requirements). Deployment validation does not replace automated tests.
+Run the applicable tests listed in [Testing Requirements](pathname://../../CONTRIBUTING.md#testing-requirements). Deployment validation does not replace automated tests.
 
 For complete validation procedures, testing templates, and cost optimization strategies, see [Deployment Validation Guide](deployment-validation.md).
 

docs/contributing/ROADMAP.md

@@ -3,7 +3,7 @@ sidebar_position: 3
 title: "Physical AI Toolchain Roadmap"
 description: "Project roadmap covering documentation, testing, CI/CD, governance, security, and OpenSSF compliance through Q1 2027."
 author: wberry
-ms.date: 2026-09-21
+ms.date: 2026-09-22
 ms.topic: reference
 keywords:
   - roadmap
@@ -32,7 +32,7 @@ Tier IDs, stage names, boundaries, and the fleet vocabulary are defined once in
 
 ## Current State
 
-The latest release recorded in [CHANGELOG.md](../../CHANGELOG.md) is v0.8.0, dated 2026-05-08; `package.json` also reports 0.8.0.
+The latest release recorded in [CHANGELOG.md](pathname://../../CHANGELOG.md) is v0.8.0, dated 2026-05-08; `package.json` also reports 0.8.0.
 Releases v0.2.0 through v0.8.0 are recorded, but release numbers alone do not establish completion of every originally planned deliverable.
 The priorities, success-metric baseline, and timeline below retain the historical planning record. OpenSSF percentages and milestone issue completion were not reassessed during this source review; future tier and autonomy plans are not marked shipped by this update.
 

docs/contributing/deployment-validation.md

@@ -3,7 +3,7 @@ sidebar_position: 8
 title: Deployment Validation Guide
 description: Validation levels, testing templates, and cost optimization for contribution testing
 author: Microsoft Robotics-AI Team
-ms.date: 2026-09-21
+ms.date: 2026-09-22
 ms.topic: how-to
 ---
 
@@ -36,7 +36,7 @@ npm run lint:sh
 npm run lint:md
 ```
 
-Run the applicable component tests as described in [Testing Requirements](../../CONTRIBUTING.md#testing-requirements). Static checks alone do not replace tests for code changes.
+Run the applicable component tests as described in [Testing Requirements](pathname://../../CONTRIBUTING.md#testing-requirements). Static checks alone do not replace tests for code changes.
 
 **When to use:** Every contribution (documentation, code, infrastructure).
 

docs/recipes/tier-0-dev/README.md

@@ -2,7 +2,7 @@
 title: "T0 — Dev: The Local Training Lifecycle Loop"
 description: Run the full capture, curate, train, validate, and run loop on one laptop and one robot with zero cloud
 author: Microsoft Robotics-AI Team
-ms.date: 2026-09-15
+ms.date: 2026-09-22
 ms.topic: tutorial
 ---
 
@@ -36,7 +36,7 @@ Everything below runs on the single machine in front of you. The only thing that
 data you copy off the robot with `cp` or `rsync`.
 
 The default instructions use plain processes. The optional
-[GPU Offload T0 plan](../../../gpu-offload/docs/05-T0-plan.md) permits a local kind
+[GPU Offload T0 plan](pathname://../../../gpu-offload/docs/05-T0-plan.md) permits a local kind
 cluster on the same laptop when process isolation, GPU allocation, or the Kubernetes
 offload controller is useful. T0 still excludes remote clusters and cloud dependencies.
 

docs/reference/workflow-templates-azureml.md

@@ -2,7 +2,7 @@
 title: Workflow Templates (AzureML)
 description: Selected AzureML workflow templates for RL, LeRobot, and SiL training and evaluation.
 author: Microsoft Robotics-AI Team
-ms.date: 2026-09-21
+ms.date: 2026-09-22
 ms.topic: reference
 keywords:
   - azureml
@@ -12,7 +12,7 @@ keywords:
   - evaluation
 ---
 
-Selected AzureML workflow templates for RL, LeRobot, and software-in-the-loop (SiL) training and evaluation, consistent with the [AzureML workflow index](../../workflows/azureml/README.md). This is not an exhaustive inventory and does not cover VLA workflows.
+Selected AzureML workflow templates for RL, LeRobot, and software-in-the-loop (SiL) training and evaluation, consistent with the [AzureML workflow index](pathname://../../workflows/azureml/README.md). This is not an exhaustive inventory and does not cover VLA workflows.
 Template names, defaults, and paths in this page are derived from the YAML files
 in `training/` and `evaluation/`.