HostAgentics Blog Publishing
Last updated: 2026-08-06
This document defines how blog content is written, validated, and published on HostAgentics. It exists because content integrity is a product decision: the blog is a source of technical orientation and honest product information, and the platform enforces content-integrity rules in code (`packages/content`, article validation). This page is the rulebook for authors, editors, and reviewers.
1. Content integrity rules (non-negotiable)
**Sourced articles.** Every factual claim about n8n, OpenClaw, or Hermes Agent must be traceable to a source. Articles carry a `sources` list (official docs, release notes, upstream repositories). Claims without sources are rewritten or cut.**Last-updated dates.** Every article carries `lastUpdated` (and the blog post record tracks `modifiedAt`). When an article is factually refreshed, the date is updated. Outdated articles are updated or unpublished — never left stale with a fresh-looking URL.**No fake testimonials.** The blog does not invent customer quotes, case studies, or "users report…" claims. No fictional customers, no fabricated personas.**No fake metrics or benchmarks.** No invented performance numbers, fake benchmark comparisons, or fabricated adoption statistics. Where a real number is cited, it must come from a cited source. Where we do not have a number, we say so.**No invented uptime or reliability claims.** No "99.9% uptime" or similar claims: HostAgentics does not yet offer a formal SLA, and the blog must not imply one. Reliability statements are limited to what the platform actually documents (daily provider snapshots, monitored health, automatic restart).**No "unlimited executions".** Plans have hard limits (see `docs/resource-limits.md`); the blog never says "unlimited" about anything that is capped — executions, tasks, sessions, storage, or transfer.**No provider identifiers.** The blog is customer-facing: the infrastructure provider is never named. The customer-facing term is "HostAgentics Cloud". (Only `docs/subprocessors.md` names the provider, and only because it is legally required.)**No speculative pricing or features.** Numbers that are not determined are marked "to be published" or described qualitatively; nothing speculative.2. Frontmatter format
Every article must carry this frontmatter (validated by `packages/content`):
```yaml
title: "Article title"
description: "One-paragraph description used in listings and meta tags."
publishedAt: 2026-08-01
lastUpdated: 2026-08-06
author: "Author name"
tags: [tag1, tag2]
sources:
- title: "Official documentation"
url: "https://..."
limitations:
- "Any limitations of the claims in this article."
```
`publishedAt` — first publication date; never backdated for SEO.`lastUpdated` — date of the last factual revision; must be current on every publish.`sources` — one or more named, linkable sources for the article's factual claims.`limitations` — explicit statements of what the article does not claim (e.g. "not legal advice", "not a benchmark", "licensing terms may change"). An article with material claims but no limitations section fails validation.3. Publishing workflow
1. **Draft** — write with the integrity rules above; include sources inline and in frontmatter. Blog posts start as drafts (`blogPosts.draft=true`); drafts are never indexed or linked from listings.
2. **Review** — an editor checks: every claim sourced, no fake metrics/testimonials, no SLA/uptime claims, no "unlimited", no provider identifiers, limitations present, `lastUpdated` current.
3. **Validation** — the content-integrity tests must pass (section 4). A failing article cannot be published.
4. **Publish** — set `publishedAt` (if not set), clear the draft flag, and schedule or publish. `modifiedAt` updates on every change thereafter.
5. **Maintain** — revisit articles when the underlying software or platform facts change; update `lastUpdated` and refresh or remove stale content.
4. Content-integrity tests
`packages/content` runs automated checks on every article (in CI and before publish):
Frontmatter completeness: required fields present, valid dates, non-empty sources and limitations where required.**Banned-pattern scan**: no "unlimited executions/tasks/sessions/storage/transfer", no invented uptime percentages, no SLA language, no fake-testimonial phrasing.**Provider-leak scan**: no provider identifiers in any article body.**Source hygiene**: `sources` entries have titles and URLs; URLs point at real, reachable domains.**Price consistency**: any article mentioning plan prices matches the canonical pricing records (`packages/pricing`) — prices in articles are checked against the source of truth.A test failure blocks publish. If a rule needs adjustment, the rule change is reviewed with the same care as the article.
5. Related documents
`docs/seo.md` — how content is surfaced to search engines`docs/ai-discoverability.md` — how AI crawlers are handled`docs/fixed-pricing.md` — the canonical numbers articles may reference`docs/subprocessors.md` — the only customer-visible document that names the infrastructure provider