#113 [P1] Add Agent Reliability Infrastructure + Architectural Verification Contracts concept pages (by 2026-06-20)

[Anonymous]

Originally created by: TheoV823

Target: ~2026-06-20 (one month from PR [#112])
Priority: P1 (next wave after P0 trio merged in [#112])

Context

P0 concepts shipped in [#112]: Agent Verification, Governance Provenance, Execution Surfaces.

Per the ontology roadmap, the next concepts on the slate are P1 [#4] and P1 [#5] below. They fit cleanly into the existing graph and unlock additional internal linking and ecosystem positioning without diluting authority. Holding them for a month-ish gives the P0 pages time to settle and gives space to see what real questions readers and crawlers raise about the P0 trio before locking the P1 framing.

Pages to add

1. Agent Reliability Infrastructure

URL: /concepts/agent-reliability-infrastructure/

Core idea: The infrastructure layers required to keep autonomous systems reliable over time — orchestration, memory, observability, governance, provenance, verification. The hub concept that names the whole category the eight-layer article maps.

Why P1: Becomes the bridge concept into broader ecosystem discussion. Potentially high-ranking. Positions Mneme inside the larger reliability story without diluting the governance/verification positioning.

Connects to: all 6 capability layers as related concepts; the emerging-ai-agent-infrastructure-stack insight as its canonical operational essay.

2. Architectural Verification Contracts

URL: decide between (a) expanding /concepts/verification-contracts/ or (b) new /concepts/architectural-verification-contracts/.

Core idea: Machine-readable constraints that define what architectural properties must remain true during autonomous execution. The compiler-analogy bridge; the deterministic-enforcement framing; the ADR → enforcement pipeline explanation.

Why P1: Strongest positioning piece against probabilistic/RAG systems. Becomes the canonical reference for the "compile architectural intent" framing in future articles.

Decision needed before writing: does this supersede or extend the existing verification-contracts page? If extend: rewrite with the architectural framing as the dominant cut. If supersede: redirect + rename.

Acceptance criteria

• [ ] Each page satisfies ADR-006 (SEO + JSON-LD + breadcrumb + mobile hamburger)
• [ ] Each page has an OG image registered in scripts/generate_og_images.py and committed
• [ ] Cross-links: concepts index cards + JSON-LD hasPart; sitemap.xml; llms.txt
• [ ] Cross-links from emerging-stack article, agent-verification, governance-provenance, execution-surfaces where natural
• [ ] Per ADR-011: no "memory" / "RAG" / "context layer" framing; runtime-stack vocabulary preserved
• [ ] Squash-merge commit title follows the taxonomy

What this issue is not

Do not expand to all P1/P2 items at once. Holding the cadence ("5 concepts immediately, then 2 more, evaluate") preserves authority. The 11–15 items on the roadmap (Architectural Provenance, Agentic Drift, Runtime Governance, Deterministic Governance, Architectural Observability, Multi-Agent Governance, Intent Preservation, Repository-Native Governance, Autonomous SDLC Governance) are deliberately deferred until P1 has shipped and been observed.

Related

• PR [#112] — P0 concept trio (Agent Verification, Governance Provenance, Execution Surfaces)
• /insights/emerging-ai-agent-infrastructure-stack/ — the originating article
• ADR-006, ADR-007, ADR-011

https://claude.ai/code/session_01BtweNjA6wcGNWqqgVVtyw9