# Themes

The load-bearing ideas across both threads.

- [The discovery-to-governance lifecycle](discovery-to-governance.md) — Intent → retrieve → select/bind → authorise → execute → record evidence.
- [Agent-ready, not agent-trusting](agent-ready-not-agent-trusting.md) — Understanding a tool is never authority to use it.
- [The missing collaborative-learning layer](missing-collaborative-learning-layer.md) — Orchestration standards don't define how institutions jointly improve a model.
- [Failure taxonomy](failure-taxonomy.md) — Seven failure classes, each with its own controls — not just 'stale data'.
- [Evaluation moved from format to execution](evaluation-format-to-execution.md) — Benchmarks shifted from argument syntax to state, policy and outcome reliability.
- [Recommendations and roadmap](recommendations-roadmap.md) — Harden the substrate in stages: inventory → control plane → institutional capability.
- [UK public-sector implications](uk-public-sector-implications.md) — Roles, SLAs and liability shift as agents move from answering to acting.