Add scenario learning proposal binding v2

[mdheller]

Summary

Fresh replay of the Memory Mesh scenario-learning proposal binding onto current main.

Supersedes #31, which diverged from main.

Adds

• schemas/scenario-learning-proposal-binding.schema.json
• examples/scenario-learning/scenario-learning-proposal-binding.example.json
• scripts/validate_scenario_learning_proposal_binding.py
• .github/workflows/scenario-learning-binding.yml

Updates

• Makefile
  • adds explicit validate-scenario-learning-binding target
  • does not add it to aggregate make validate
• README.md
  • documents scenario-learning proposal bindings

Boundary

No durable writeback is enabled or performed.

The binding enforces:

• proposalMode = review_only
• pending human review
• runtime decision receipt references
• evidence references
• policy decision references
• no raw sensitive payload storage
• memory effect remains pending_review or blocked
• writeback.enabled = false
• writeback.performed = false
• writebackRef = null

Relationship to SCOPE-D / Ontogenesis

This is the Memory Mesh companion to the SCOPE-D adversarial scenario contract and Ontogenesis adversarial scenario boundary documentation. It gives scenario-derived lessons somewhere to land without turning scenario output, ATT&CK coverage, dashboard exports, or model summaries into durable memory by accident.

Validation

Dedicated workflow: scenario-learning-binding.

Local target: make validate-scenario-learning-binding.