Macro-3 Channel Agent-API Final Case — R2-B2 (2026-06-19)
Macro-3 Channel Agent-API Final Case — R2-B2 (2026-06-19)
Date: 2026-06-19 · Workstream: R2-B2-MACRO-3-OPTION-D-OWNER-DELEGATE-DECISION-PACKAGE-2026-06-19 (Deliverable 8 of 60) · Editorial revision: rev1
Class: channel case (agent-api) · READ-ONLY · NON-ENACTING · NON-AUTHORIZING · CHANNEL_AUTHORITY_DRIFT guard · NO channel selected · NO write performed.
Metadata convention. Editorial revision (rev1) only. Storage revision/
content_lengthauthoritative at read time.
0. Status and non-authorization
STATUS: PASS — engineering / decision-prep. The strongest honest case for agent-api, without selecting it. Engineering PASS ≠ authority PASS. Default: HOLD.
1. Purpose
Give the Owner the best argument for the agent-api executor so the channel decision is informed.
2. Sources / evidence read
Macro-2 agent-api decision memo, weighted comparison; FRESH list_docker (:8090 healthy), dot_agent_api_contract=2/0-birth-bound, switches OFF. Main process, no reader-agents.
3. Accepted baseline (carried)
agent-api = candidate, NOT selected; governance-aligned (contract-bound, observable); DRY_RUN-gated.
4. Evidence / analysis
For agent-api: (a) governance traceability — runs are contract-bound and observable, so a birth-inspection run leaves an auditable trail; (b) already healthy — container Up 2 weeks on :8090; (c) DRY_RUN gate exists — dry_run_only=true means a contract can be exercised without enacting; (d) aligns with the existing DOT/contract governance model rather than bolting on an out-of-band cron. Best when the Owner values observability / contract-bound promotion / auditability. This is the lean if the Owner forces a single pick (per D6), but approval remains required.
5. Recommendation / matrix / result
Recommend agent-api if the Owner prioritizes observability / auditability, and as the forced-pick lean for Macro-4 readiness. Not selected here. Pair with the agent-api risk map (D10) and proof plan (D12).
6. Owner-gated future work
Binding/promoting an agent-api birth-inspection contract (even DRY_RUN) is Owner-gated; forbidden now.
7. What remains unresolved
0 birth-bound contracts today; switches OFF; the contract would have to be designed + DRY_RUN-proven first.
8. Ready for GPT/Codex review
Yes — Codex should attack whether "governance-aligned" justifies the larger surface vs host cron's lower blast radius.