seo-guest-posting— agent skill

Overview

Guest posting still works in 2026 — but only as editorial contribution. Google's link spam policy names "links with optimized anchor text in articles, guest posts, or press releases distributed on other sites" and paid posts that pass ranking credit; its scaled content abuse policy covers mass AI-written placements; the August 2025 spam update and the 2026 spam updates (March, June, August) kept enforcing both. Volume is a liability. "Niche edits" sold as guest-post alternatives carry the same risk as paid guest posts. The tactic survives only when each post is editorially earned, topically aligned, written to publication standard, and linked back with diverse natural anchors.

Third-party readings of the March 2026 (Mar 27–Apr 8) and May 2026 core updates saw aggregation-model sites lose; Google gave no specifics. Implication: a guest post on an aggregator (low first-hand content, mostly link-out) passes less authority than the same post on a primary-source publication in the niche — pair qualification with a "primary source or intermediary?" call. Google's site reputation policy also targets third-party content published on a strong host mainly to borrow its ranking signals (manual actions no longer apply to EEA searchers since Aug 30, 2026). Topical fit is no longer optional.

This skill encodes the 7-stage workflow: map your authority to publication tiers, discover targets, qualify each against an 11-point rubric, pitch a specific content gap (not "a guest post"), write to the host's standard, link without over-optimizing, and follow up toward a multi-post relationship. Use BEFORE writing any pitch email. The deliverable is a strategy + brief; downstream copywriting belongs to the agent layer.

Instructions

Core Rules

• Pitch a content gap, not "a guest post." Editors get 50+ generic pitches per week and delete them in seconds. Accepted pitches name a topic the publication should already cover and offer to fill it. "I'd love to contribute a guest post" is the kiss of death.
• Match your authority to the publication tier. Tier-1 wants named expertise and prior tier-1 placements; with no portfolio, start lower and build up. Tier-mapping (Stage 1) is non-negotiable.
• Anchor diversification is a practitioner heuristic, not a Google ratio: branded majority (~40–50%), partial-match 20–30%, semantic/long-tail 15–25%, generic 5–10%, exact-match under 10%. Google's policy targets optimized anchors in guest posts, so default to branded. Track distribution across all placed posts.
• 2–4 posts/month is a sensible cadence ceiling. A sudden run of placements reads as a campaign footprint.
• Body links beat bio links. 1-2 contextual body links + 1 author bio link is the safe default; body placement earns more referral traffic. If the link doesn't help the reader at that exact point, drop it.
• Never pay for a placement that passes ranking signal. Paid + no rel="sponsored" = penalty. "Vetted networks" charging $200-500 per placement are usually disguised PBNs.
• Each post must be unique. Duplicates get filtered, burn editor trust, and mass republishing reads as scraping. Re-pitching a topic to multiple sites is OK only if each post is independently written.
• One pitch, one topic-set. Don't blast the same 3-topic pitch to 50 sites; research each publication, reference its actual content, fill its specific gap.
• AI-written drafts are a liability. Mass AI content falls under the scaled content abuse policy, and editors reject it. AI for outlining/research support is fine; ship human-written prose at publication standard.
• Output a brief, not a finished email. This skill produces strategy, target list, topics, and pitch architecture; email and post writing are owned elsewhere.

The 7-Stage Workflow

#### Stage 1 — Tier-Mapping (where YOU can realistically pitch)

Before discovering any target, make the user honestly map their authority. Pitching above your tier is the #1 cause of pitch failure.

Author Authority Inventory:

1. Prior published work — list it. URLs, publication, date.
2. Named expertise — credentials, role, years in the field, specific niche depth.
3. Original assets — proprietary data, original research, named tools, signature frameworks.
4. Social proof — speaking engagements, podcast appearances, citations, follower count if relevant.
5. Domain you'd link from — the brand/site doing the pitching. Its DR, niche relevance, recognizable name.

Tier Mapping:

The honest rule: with zero prior placements, START at T4 and ladder up. A tier-1 pitch with no portfolio wastes 4 weeks on unanswered emails.

Output of Stage 1: a written tier mandate. Example: "Author has 2 T3 placements + proprietary benchmark data. Realistic mandate: T2 (stretch) and T3 (core), no T1 attempts this quarter."

#### Stage 2 — Discovery (operators + competitor reverse-engineering)

PARALLEL-FIRST: fire all discovery searches in ONE block.

Operator library (substitute [topic] with the niche term):

[topic] intitle:"write for us"
[topic] intitle:"write for me"
[topic] intitle:"contribute to"
[topic] intitle:"contributor guidelines"
[topic] inurl:/guest-post/
[topic] inurl:write-for-us
[topic] inurl:contribute
[topic] "submit a guest post"
[topic] "submit an article"
[topic] "now accepting guest posts"
[topic] "guest writer wanted"
[topic] "become a contributor"
[topic] ("write for us" OR "guest post" OR "contribute" OR "submit")

Competitor reverse-engineering (the highest-yield discovery method):

• Take a competitor URL → free Ahrefs Backlink Checker → top 100 referring domains
• Filter referring page URL or anchor text for "guest", "contributor", "by [author name]"
• Each hit is a publication that already publishes content in your niche AND accepts outside contributions

Topic-research / content-gap tools:

• For each candidate publication, identify topics in your niche they haven't covered (use site search + keyword gap analysis if available)
• This becomes raw material for Stage 4 (topic ideation)

Niche-specific operator variants:

Pitfall: openly advertised "write for us" pages attract paid networks. The best publications often have none — reach them via competitor reverse-engineering and direct editor relationships.

#### Stage 3 — Qualify Each Target (the 11-point rubric)

Walk top-to-bottom, kill on first failure. This is a kill-rule list, not a score.

A prospect that survives all 11 → eligible for pitch. Anything else → skip.

Network-graph check (for advanced users with paid tools): if Ahrefs/Semrush show backlinks from this domain "negatively impact" linked sites, that's a hard strike — the site is a known toxic linker.

#### Stage 4 — Topic Ideation (per qualified target)

This stage separates ignored pitches from accepted ones: generic topic offers fail, specific content-gap offers convert.

Per-target ideation process:

1. Inventory the publication's recent content — last 30 posts. Group by sub-topic.
2. Identify the gaps — what sub-topics in their declared niche are missing or under-covered? What perspectives haven't they shown?
3. Match to the user's unique assets — proprietary data, lived experience, technical specialty, contrarian framing.
4. Generate 3–5 topic candidates that:

• Fit the publication's editorial niche exactly
• Don't duplicate something they ran in the last 6 months
• Bring something only the user can bring (data, case study, technical depth)
• Match a known search-intent pattern in the niche (keyword research, if needed, is handled outside this skill)

1. Pre-write the value proposition — for each topic, one sentence on why their audience needs this.

Topic angles that work:

Topic angles that fail:

• Generic listicles ("10 ways to improve your SEO")
• Broad topics already covered ("What is content marketing?")
• Topics that promote your product (editors smell this in 5 seconds)
• "Trends in [niche] for 2026" pieces (every publication has 5 of these)
• Anything resembling a pitch deck

Output of Stage 4: for each qualified target, 3–5 specific topic offers, each one sentence + a value prop.

#### Stages 5–8 — Pitch, Post, Links, Follow-up

Quick guidance (full templates in the reference below):

• Stage 5 (Pitch): 80–150 words. Subject references their content, body = personalization hook → credibility line → 3 topic offers → soft CTA. Personalized pitches get several times the replies of generic ones.
• Stage 6 (Post): word count and depth match the tier (T1: 2.5–4k words + original data; T2: 1.5–2.5k; T3: 1–1.5k; T4: 0.8–1.2k). Universal rules: original, human-written, match publication voice, no promotional content in body.
• Stage 7 (Links): 1–2 body links + 1 bio link. Anchor heuristic: branded majority, exact-match under 10%; default to branded on controlled placements.
• Stage 8 (Follow-up): 7-day bump, 14-day final, thank-you after publish, re-pitch at 90 days. Goal: 3–4 posts in 12 months, not one-offs.

Full templates (subject formulas, pitch architecture, tier-by-tier writing standards, anchor-distribution table, link-placement rules, follow-up cadence): see reference/detailed-stages.md.

Decision guide and Google's policy

Edge cases (paid networks, tier-jumping, ghostwriting, syndication, exact-match push-back) and Google's exact rules on what's editorial vs violation: see reference/detailed-stages.md. The intent test stands: would this guest post exist if SEO didn't exist? If yes → editorial. If no → violation in waiting.

Tool inventory (free + paid) for discovery, qualification, and outreach: see reference/detailed-stages.md.

Examples

Example 1: Tier-3 placement, dev-tools SaaS

Setup:

• Author: senior platform engineer, brand: CI/CD SaaS
• Portfolio: 1 prior post on a tier-3 dev blog
• Target: a DR-45 dev-ops focused publication, ~25k organic monthly, named editor

Stage 1 — Tier mandate: T3 core (matches portfolio); T2 stretch with an exceptional pitch.

Stage 2 — Discovery operators (parallel):

"continuous integration" intitle:"write for us"
"devops" intitle:"contribute"
"ci/cd" inurl:/contributors/
site:[competitor.com] "guest post" → reverse-engineer competitor's link sources

Stage 3 — Qualification: topical fit ✅ (last 30 posts: 60% CI/CD, 30% dev-ops, 10% adjacent); traffic ✅ (Similarweb 22k/mo organic); editorial ✅ (named editor, masthead, reject rate in guidelines); indexed/ranking ✅; outbound ✅ (GitHub, AWS docs, dev-blog peers); author diversity ✅ (50+ distinct contributors); no PBN tells ✅; free editorial, no paid-placement signals. Verdict: proceed.

Stage 4 — Topic ideation (3 candidates):

• Read last 30 posts. Identified gaps:

1. "What 18 months of monorepo CI metrics taught us about test parallelization" — original benchmark data from author's company, fills a gap (publication has covered monorepo strategy but not parallelization specifics)
2. "Self-hosted runners vs. managed CI: a 6-month cost-and-reliability comparison" — case study with numbers; publication has discussed runners but not done a head-to-head
3. "Why we shipped a CI cancellation policy and what we measured after" — process post; ties to author's company experience

Stage 5 — Pitch brief:

Subject: Loved the runner-sizing piece — pitch on monorepo CI metrics?

Hi Sarah,

Your post on right-sizing runners last month nailed the cost angle, but the comment thread kept asking about test parallelization specifically — there's a real gap there.

I'm a platform engineer at Acme Deploy. We've been instrumenting our monorepo CI for 18 months across ~3,400 daily runs and have data I haven't seen anywhere else. I wrote a tier-3 piece on runner ergonomics for [other-publication] last quarter (link).

Three angles I could write that fit your editorial scope:

  • What 18 months of monorepo CI metrics taught us about test parallelization — original benchmark data, the 4 metrics that actually matter
  • Self-hosted runners vs. managed CI: a 6-month cost-and-reliability comparison — head-to-head case study with numbers
  • Why we shipped a CI cancellation policy and what we measured after — process post on a non-obvious lever

Happy to send a detailed outline if any of these fit.

— Author Name
acme-deploy.com

Stage 6 — Post brief: tier-3 standard, 1,500 words, case study format, 3 charts from internal data, 2 internal links to the host's prior CI/CD posts. Outline only; prose is written outside this skill.

Stage 7 — Link strategy:

• 1 body link: "...we benchmarked this against the [Acme Deploy CI runner pool]..." — partial-match anchor
• 1 author bio link: "Author leads platform engineering at [Acme Deploy]." — branded anchor
• Anchor distribution check: this is post #3 for the brand. Across all 3: 2 branded, 1 partial-match. ✅ within target distribution.

Stage 8 — Follow-up: post published; share on Twitter and LinkedIn tagging publication; thank-you email to editor 1 week post-publication; re-pitch in 90 days with the runner comparison topic.

Example 2: Refusing a tier-1 reach

User: "I have no portfolio but want to pitch a guest post to TechCrunch."

Response:

Stop. TechCrunch's acceptance rate for cold pitches with no prior tier-1 placements is functionally zero. Three weeks on that pitch costs you the 5 tier-3 placements you could land in the same time.

>

Honest path: 1. Months 1–3: tier-4 (DR 10–30 niche blogs, owner-edited). Land 3–4 real bylines. 2. Months 3–6: ladder to tier-3 with the tier-4 portfolio in your credibility line. Land 2–3. 3. Months 6–12: at tier-2 with 5–7 prior placements, acceptance jumps to 15–20%. Land 2–3. 4. Year 2: with tier-2 placements, one carrying original data, a TechCrunch pitch lands at maybe 5–10% instead of 0%.

>

Pitching above your tier costs you the tier you could be building.

Do not soften.

Example 3: Refusing a paid network

User: "I found a service offering 10 guest posts per month on DR 50+ sites for $1,500 total. They handle everything."

Response:

Skip. The math gives it away:

>

- $150 per "DR 50+" placement is below the cost of even outsourced content writing (~$200-400 per post). Real editorial publishers spend money reviewing and editing — they neither pay you to publish nor accept $150 placements. - At 10/month from one provider, every network site is reused across clients: same bylines, anchor patterns, template-style content. That is the paid-link plus scaled-content footprint Google's spam policies name. - Even if the links pass initially, SpamBrain tends to neutralize them. Worst case: a manual action on your domain, not theirs, with a 6-month recovery curve.

>

Net: ~$1,500 spend, ~0 long-term link value. Real editorial costs more time but compounds — 2 earned tier-3 posts per month beat 10 paid network posts every quarter.

References

• Google Search Spam Policies — link spam definitions
• Backlinko — The Definitive Guide to Guest Blogging — Tier framework, qualification, pitch structure
• Search Engine Journal — Google penalties on guest post articles
• Search Engine Land — Guest post outreach 2026 process
• Hunter — 15 best-performing guest post email templates — pitch performance data
• Google — Optimizing for generative AI features (May 2026) — scaled content abuse, inauthentic mentions