seo-guest-posting— agent skill

Overview

Guest posting still works in 2026 — but the rules changed sharply. The October 2025 Google spam update added "AI-generated guest post farms" as a distinct violation category, targeting large-scale operations that publish thin machine content to embed paid backlinks. Volume is now actively penalized. Repetitive exact-match anchors across posts trigger manual review. "Niche edits" sold as guest-post alternatives carry the same penalty 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.

The March 2026 core update (March 27 → April 8) sharpened this further: 80% of top-3 results shifted, and Google explicitly devalued intermediary/aggregator pages in favor of primary sources. Implication for guest posting: a guest post on a site that is itself an aggregator (low first-hand content, mostly link-out) now passes far less authority than the same post on a primary-source publication in the niche. Always check whether the target site is a primary source — qualification stage 11 (penalty history) below should now be paired with a "is this a primary source or an intermediary?" judgment call. Also note: Site Reputation Abuse enforcement became fully algorithmic in the August 2025 Spam Update — placing a topically independent piece on a high-DA host now decays in 6–8 weeks (down from ~9 months under the old manual-only policy). Topical fit is no longer optional.

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

Instructions

Core Rules

• Pitch a content gap, not "a guest post." Editors get 50+ generic guest-post pitches per week and delete them in seconds. The pitches that get accepted name a specific topic the publication should already have covered and offer to fill it. The phrase "I'd love to contribute a guest post" is the kiss of death.
• Match your authority to the publication tier. Pitching tier-1 with no portfolio is a waste — they want named expertise and prior tier-1 placements. Start where you are; build up. The tier-mapping step (Stage 1) is non-negotiable.
• Anchor diversification is data-backed, not a vibe. 40-50% branded, 20-30% partial-match, 15-25% semantic/long-tail, 5-10% generic, <10% exact-match. Diversification reduces manual-action risk by 68%. Track distribution across all guest posts placed.
• 2–4 posts/month is the cadence ceiling. More than that triggers velocity-pattern detection. Quality beats quantity now in a way that wasn't true 3 years ago.
• Body links beat bio links. 1-2 contextual body links + 1 author bio link is the safe default. Body placement generates up to 387% more referral traffic than bio-only. The link must genuinely help the reader at that exact point — if it doesn't, drop it.
• Never pay for a guest post placement that passes ranking signal. Paid + no rel="sponsored" = penalty. "Vetted networks" charging $200-500 per placement are usually disguised PBNs. Real editorial requires real outreach.
• Each post must be unique. Duplicate content across publications is its own violation category. Re-pitching the same topic to multiple sites = OK only if each post is independently written.
• One pitch, one topic-set. Don't blast the same 3-topic pitch to 50 sites. Each pitch is researched per-publication, references their actual content, fills a specific gap they have.
• AI-written drafts are detectable and devalued. The Oct 2025 update specifically targets thin AI content farms. Use AI for outlining/research support if needed, but ship human-written prose at publication standard.
• Output a brief, not a finished email. This skill produces the strategy, target list, topics, and pitch architecture. The actual email and post writing are downstream concerns owned elsewhere.

The 7-Stage Workflow

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

Before discovering any target, force the user to 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: if the user has zero prior placements, START at T4, ladder up. Trying to land a tier-1 pitch with no portfolio wastes 4 weeks and generates a pile of 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: "Write for us" pages openly advertised attract paid networks. The highest-quality publications often DON'T have "write for us" pages — discovery there requires competitor reverse-engineering and direct editor relationships, not operator searches.

#### 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 is the stage that separates 6% pitch reply rates from 18%. 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

Stages 5–8 are workflow execution and have detailed templates, tier-by-tier writing standards, anchor distribution targets, and follow-up cadence. Quick guidance:

• Stage 5 (Pitch): 80–150 words. Subject references their content, body = personalization hook → credibility line → 3 topic offers → soft CTA. Personalized pitch yields 18% reply vs 6% generic.
• 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 mix: 40–50% branded, 20–30% partial-match, 15–25% semantic, 5–10% generic, <10% exact-match. Diversification reduces manual-action risk by 68%.
• 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 of target:

• Topical fit: ✅ (last 30 posts: 60% CI/CD, 30% dev-ops, 10% adjacent)
• Real traffic: ✅ (Similarweb: 22k/mo organic)
• Editorial: ✅ (named editor, masthead, reject rate referenced in their guidelines)
• Indexed/ranking: ✅
• Outbound profile: ✅ (links to GitHub, AWS docs, dev-blog peers)
• Author diversity: ✅ (50+ contributors, all distinct)
• No PBN tells: ✅
• Cost: free editorial; no paid placement signaling
• 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 writing brief:

• Tier-3 standard: 1,500 words; case study format; 3 charts from internal data; 2 internal links to host site's prior CI/CD posts
• Outline produced; prose generation handled 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. Spending 3 weeks crafting a TechCrunch pitch with no portfolio costs you the 5 tier-3 placements you could land in the same time.

>

Honest path: 1. Months 1–3: target tier-4 (DR 10–30 niche blogs, owner-edited). Land 3–4 placements. These are achievable and they're real bylines. 2. Months 3–6: ladder to tier-3 with the tier-4 portfolio in your credibility line. Land 2–3 here. 3. Months 6–12: at tier-2 with 5–7 prior placements, acceptance rates jump to 15–20%. Land 2–3 here. 4. Year 2: at this point a TechCrunch pitch has a chance — you have tier-2 placements, one with original data they'd notice. Now the pitch lands at maybe 5–10% probability instead of 0%.

>

The path matters more than the destination. 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 away what's actually happening:

>

- $150 per placement on a "DR 50+" site is below the editorial cost of even outsourced content writing (which is ~$200-400 per post). Real editorial guest posts cost the publisher money to review and edit — they don't pay you to publish, they don't accept $150 placements either. - At scale (10/month from one provider), every site in the network is reused across clients. Same author bylines, same anchor patterns, same template-style content. This is the textbook footprint the Oct 2025 update flagged as "AI-generated guest post farms." - Even if the links pass initially, devaluation lands within weeks. Worst case: manual action on the receiving domain — your site, not theirs.

>

The honest math: 10 placements at $150 each from a network = ~$1,500 in spend, ~0 long-term link value, and a manual-action risk on a 6-month recovery curve.

>

Real editorial costs more in time, but the placements compound. 2 tier-3 editorial guest posts per month, earned through the workflow above, 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
• Anchor distribution data — 68% penalty reduction
• Editorial links are the new brand mentions for AI search — 2026 framing