Skip to content

coreyhaines31/marketingskills / skills/link-building

link-building

Verified

When the user wants backlinks and mentions that lift rankings and AI citations, or asks how to get sites to link, whether a link tactic is safe, or where to buy placements: link gaps, unlinked mentions, broken links, guest posts, link marketplaces. On paid links, state Google's rule (rel=sponsored or nofollow) and the risk, then let the user decide. For press, see public-relations. For directories, see directory-submissions. For reviews, see referrals. For communities, see community-marketing.

FileKindSize
SKILL.mdskill md15.7 KB
evals/evals.jsonconfig7.2 KB
references/prospecting-and-outreach.mdmarkdown4.6 KB
references/risk-policy.mdmarkdown6.4 KB
references/tactics.mdmarkdown7.0 KB

evals/evals.json

View on GitHub
{
  "skill_name": "link-building",
  "evals": [
    {
      "id": 1,
      "prompt": "Our pricing page is stuck at position 9 for 'invoicing software for freelancers'. Help me build backlinks to it.",
      "expected_output": "Should check for product-marketing.md first. Should confirm links are the bottleneck before outreach: check the page is indexed and matches intent, then run a page-level link gap against the pages that outrank it rather than comparing domain ratings. Should recommend internal links from strong pages to the pricing page. Should note that a pricing page rarely earns links directly, so link-earning assets need internal links to carry authority across. Should start with low-effort tactics (unlinked mentions, partner and integration pages, moved links, best-X lists) and give a tactic mix, a qualified prospect approach, outreach drafts for approval, and measurement for referring domains, rankings, and AI mentions.",
      "assertions": [
        "Checks for product-marketing.md",
        "Runs a page-level link gap before recommending outreach",
        "Checks the page itself (intent, indexing) before blaming links",
        "Recommends internal links to the target page",
        "Starts with low-effort tactics such as unlinked mention reclamation",
        "Drafts outreach for approval rather than sending it",
        "Measures referring domains to the target page and rankings"
      ],
      "files": []
    },
    {
      "id": 2,
      "prompt": "An agency offered us 20 do-follow guest post links on DR 60+ sites for $3,000. Good deal?",
      "expected_output": "Should identify this as buying links that pass ranking credit, which Google's spam policies name as link spam, and state the risk once: Google may ignore the links, and a large or obvious pattern can bring a manual action. Should leave the decision to the user rather than refusing. Should explain that paid placements are inside the policy only with rel=sponsored or nofollow and disclosure, and that keyword anchors in guest posts are a named example. Should give the vendor-vetting checks (price per followed link, sites that publish anything, chosen anchors, reporting by link count) and the per-site checks to run if the user goes ahead (topic fit, real traffic, outbound links, anchors, pace). Should not treat DR 60+ as proof of quality or promise a ranking lift. Should name alternatives: earned outreach, digital PR, sponsored placements valued for traffic and AI visibility.",
      "assertions": [
        "States that paid followed links are outside Google's policy and names the risk",
        "Leaves the decision to the user instead of refusing",
        "Cites Google's link spam policy and the sponsored or nofollow requirement",
        "Does not treat domain rating as a reason to buy",
        "Gives vendor-vetting signals",
        "Gives per-site checks for going ahead",
        "Offers alternatives"
      ],
      "files": []
    },
    {
      "id": 3,
      "prompt": "ChatGPT never mentions us when people ask for the best project management tool for agencies. How do we get mentioned on the sites it pulls from?",
      "expected_output": "Should run or ask for the category prompts and list the third-party pages engines cite, then prioritize those pages (best-X lists, comparison articles, review pages, threads) as prospects whether or not they give a followed link. Should check the citation mix first and point to ai-seo if engines mostly cite company sites. Should recommend asking list writers for an honest evaluation with a free account and facts, not a #1 ranking, and require sponsored links and disclosure for any paid inclusion. Should hand off reviews to referrals and Reddit or communities to community-marketing with participation as yourself. Should measure mentions on cited pages and AI mention rate.",
      "assertions": [
        "Builds the prospect list from pages AI engines cite for the prompts",
        "Counts mentions without links as wins for AI visibility",
        "References the citation mix or ai-seo",
        "Asks for honest evaluation rather than a #1 ranking",
        "Requires sponsored link and disclosure for paid list inclusion",
        "Does not recommend planted Reddit posts or sockpuppets"
      ],
      "files": []
    },
    {
      "id": 4,
      "prompt": "Ahrefs says we have 340 toxic backlinks. Should I disavow them all?",
      "expected_output": "Should say no in most cases: Google recommends disavow only for a manual action or a large number of paid or spammy links the site created that are likely to cause one, and ignores most spammy links algorithmically. Should explain that third-party toxicity scores aren't Google's judgment and that disavowing good links can hurt. Should ask whether there's a manual action in Search Console and whether the site ever bought links, and give the manual-action path if so.",
      "assertions": [
        "Does not recommend disavowing all flagged links",
        "Explains when Google recommends the disavow tool",
        "Asks about a manual action or a history of bought links",
        "Notes toxicity scores are third-party estimates"
      ],
      "files": []
    },
    {
      "id": 5,
      "prompt": "Write an outreach email asking a blogger to swap links with us. We'll link to them from our partners page if they link to our homepage.",
      "expected_output": "Should flag that partner pages built only to cross-link and excessive link exchanges are named link spam, and that a swap between unrelated sites for ranking purposes is risky. Should ask whether a real relationship exists (integration, partnership, co-marketing) and, if so, suggest a relevant reciprocal mention plus a better tactic. Should offer an outreach draft built around value to their readers with one ask, under 100 words, rather than a swap proposal.",
      "assertions": [
        "Flags link exchanges for ranking as a spam risk",
        "Asks whether a genuine partnership exists",
        "Offers a compliant alternative outreach draft",
        "Keeps the draft short with one ask"
      ],
      "files": []
    },
    {
      "id": 6,
      "prompt": "We have about $2,000 a month for links and nobody has time for outreach. Which marketplaces or services should we look at?",
      "expected_output": "Should name more than one option and separate the two routes: a managed placement service that does the outreach, and self-serve marketplaces where the user picks each site and price. Should describe what each named tool is in a clause, without calling any of them the best or recommending one over the rest. Should say plainly that placements bought this way are paid links, give Google's rule (sponsored or nofollow) and the risk once, and leave the decision to the user. Should give the per-site checks (topic fit, real traffic, outbound links, anchors, pace) and say they apply to each site, not to the marketplace. Should not promise a ranking lift.",
      "assertions": [
        "Names more than one tool or service",
        "Separates managed services from self-serve marketplaces",
        "Does not endorse one tool as best",
        "States that these are paid links, with Google's rule and the risk",
        "Gives per-site checks",
        "Does not promise a ranking lift"
      ],
      "files": []
    }
  ]
}
receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementationStars: 322kUpdated Claude CodeCodexVerifiedInstall
requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirementsStars: 322kUpdated Claude CodeCodexVerifiedInstall
using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallbackStars: 322kUpdated Claude CodeCodexVerifiedInstall
verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions alwaysStars: 322kUpdated Claude CodeCodexVerifiedInstall