optimize: mcp-plus-pilot-tools-and-network — title/meta, FAQ, internal links

[pstayets]

CTR harvest for /blog/mcp-plus-pilot-tools-and-network: 218 impressions, 0 clicks, position 6.8.

Changes: front-loaded the primary query (pilot mcp / agent-to-agent communication) in the title and rewrote the meta description to 154 chars; added a 4-item FAQ block (renders FAQPage JSON-LD); added id attributes to the H2s so the page TOC renders; added contextual internal links to same-cluster posts (connecting-mcp-servers-across-agents, ai-agent-app-store); removed a stale third-party download figure to stay within the no-figures brand guardrail; added the page to llms.txt.

[pstayets]

@TeoSlayer @Alexgodoroja friendly ping — the SEO PR queue has stacked up since the 13th: 21 PRs open, all green (CI passing, no conflicts, DCO signed), oldest is this one from 07-31. None need code review beyond a sanity pass — they're title/meta/FAQ optimizes and standalone /learn pages.

If you only have time for a few, these six carry most of the measured impact:

1. optimize: /blog/direct-communication-protocols-ai-agents-guide — head-term title/meta, FAQ, interlinks #200 — head-term fix for 'agent communication protocol', our biggest non-brand query (451 imp/28d, stuck pos 23)
2. site: add /agents.txt + /.well-known/integrations.json agent-discovery manifests #198 — /agents.txt + /.well-known/integrations.json agent-discovery manifests (GEO answer-share is still 0/14; these are the surfaces assistants check)
3. blog: restore IETF Internet-Draft page — 2026 reframe, interlinks #213 — restores the IETF draft page that was deleted in the 07-31 sweep and still pulls ~650 imp/28d while 404ing
4. optimize: /blog/pilot-vs-tcp-grpc-nats-comparison — NATS vs gRPC title/meta, FAQ, links, guardrail sweep #201 — 'nats vs grpc' ranks pos 7 with zero clicks; title/meta fix on the serving page
5. learn: How to Set Up a Persistent, Stable Webhook Endpoint URL That Doesn't Change on Restart #191 — the persistent-webhook question ranks pos 4-5 with zero clicks; this learn page answers it verbatim
6. blog: Headscale vs Nebula vs Tailscale — expand mesh-overlay comparison #199 — extends the mesh-comparison family, the only cluster currently earning non-brand clicks

The rest (#190, #195, #202, #204-#206, #208-#212, #214-#216) are the same class of change and can batch-merge anytime. I'll rebase anything that goes stale. Thanks!

[TeoSlayer]

Closing the conflicting branch after product review. The composition topic is useful, but this version has unsupported scale/task-delegation claims, a broken Python example, and an overlong title. Reintroduce the intent as a focused current-main article with executable examples.