[Parent] Track agent runtime and execution pipeline improvements

Umbrella issue tracking enhancements to how agents are invoked, how prompts are assembled at runtime, and how execution context is surfaced.

## Related issues

- #1486 — adopt Copilot sidecar (SDK-driver) execution model
- #860 — built-in PR context precomputation for PR-triggered agents
- #432 — inject runtime parameter values into the agent prompt via ado-script
- #237 — decouple prompt body from compiled pipeline output (runtime-import approach)
- #1375 — reject unwireable cross-job provider-env macros on engine.env

> Generated by [Issue Arborist](https://github.com/githubnext/ado-aw/actions/runs/30043084056) · 51.3 AIC · ⌖ 6.31 AIC · ⊞ 6.1K · [◷](https://github.com/search?q=repo%3Agithubnext%2Fado-aw+is%3Aissue+%22gh-aw-workflow-call-id%3A+githubnext%2Fado-aw%2Fissue-arborist%22&type=issues)

This is an umbrella issue rather than a self-contained change. Start by reviewing the related issues #1486, #860, #432, #237, and #1375, then select a specific enhancement with its own scope and entry points. Done means the chosen runtime or execution-pipeline improvement is implemented and its associated issue has clear completion evidence.

Written by the indexing model from the issue text.