Pattern Proposal: Steering Runtime Projection

Summary

Compiles architectural deltas and capability records into a structured runtime index that projects contextual PromptFacts into agent environments.

Candidate Pattern

• ID: steering-runtime-projection
• Classification: governance
• Maturity: established
• Type: implemented

Observed Signals

No specific signals listed.

Evidence

Supporting Evidence

• src/cas_runner/steering/models.py (SteeringRuntimeIndex, PromptFact, CapabilitySet, ArchitecturalDelta) - Defines runtime data models for projecting capability facts into agent steering context.
• .github/workflows/cas-build-steering-projection.yml - Automated workflow projecting steering facts from architecture deltas.

Scope Assessment

• Evidence Strength: high
• Confidence: high
• Applies to: ideas-to-life-cas

Rationale

• Intent: Ensure AI agents operate within architectural boundaries by projecting exact governance constraints and delta context into runtime prompts.
• Context: Agent execution across repositories requiring strict compliance with architectural decision records and evolving capability models.
• Force: Full architecture specifications exceed context window limits and introduce irrelevant noise
• Force: Agents without explicit steering violate enterprise architectural standards and conventions

Traceability

Artifact source: proposed-patterns