Define Org-Wide Agent Run Contract

## Goal
Define one mandatory execution contract for agent-driven changes across repositories.

## Why
Developers need a concrete, repeatable checklist for what must run before merge, independent of optional proprietary accelerators.

## Mandatory Contract (Applies to all repos)
1. **Lane A (mandatory)**: all merge-critical checks run with open-source tooling only. Merge eligibility is determined exclusively from Lane A results.
2. **Lane B (optional)**: proprietary or LLM-assisted accelerators may improve productivity but cannot be required for merge and cannot override a Lane A denial.

## Required Contract Fields
- Runtime/container requirements (image, tool versions, reproducibility settings)
- Required command sequence (build, tests, policy checks, artifact generation)
- Required evidence artifacts and retention
- Failure categories and retry strategy
- Waiver process (approver, reason, expiry, audit reference)
- Fallback behavior when optional accelerators are unavailable

## Implementation Tasks
- Publish a reusable contract template (`run_contract.yaml` or equivalent) for all repos.
- Define a per-repo command mapping table.
- Define rerun consistency criteria and tolerance window.
- Define branch protection check names linked to contract stages.

## Docs-as-code pilot (first implementation)
- Build docs and generate needs data
- Generate traceability metrics JSON via existing traceability tooling
- Enforce thresholds via existing traceability gate
- Run repository tests
- Emit standardized evidence artifacts

## Done When
- Contract template is approved by maintainers and referenced by onboarding issues.
- At least one pilot repo (docs-as-code) passes all required Lane A checks end-to-end.
- Contract includes explicit fallback behavior for unavailable accelerators.

Parent: eclipse-score/score#2852