# anchor-policy: a Claude Code skill by Scalably

Canonical: https://scalably.io/skills/anchor-policy
Source: https://github.com/scalably-io/agent-skills/blob/main/skills/anchor-policy/SKILL.md (MIT, v1.0.0)
Plugin: seo-ops · Provenance: derived from a Scalably production skill, source commit f0f1753, verified 2026-09-07
Integrity: sha256 of the raw SKILL.md served at https://scalably.io/skills/anchor-policy.md is 354b7ae365226c0d3316b7558205c1cd82354b522ccfad4d290be2ed21679f5b. Verify: curl -s https://scalably.io/skills/anchor-policy.md | shasum -a 256
This is the machine-readable representation of the page at the canonical URL. Same facts, denser format. The raw SKILL.md is the installable artifact.

## Direct answer
Validate anchor text placement, URL correctness, and SEO positioning in a guest-post draft against a target anchor spec: checks that anchors land in the correct sections, the primary anchor is the first external link, the introduction is link-free, and spacing rules are met, for both the Standard/PR and Listicle anchor models. Triggers: check anchors, validate links, anchor policy, anchor compliance, verify anchor text, anchor QA, link placement check.

## Install
- Claude Code: `/plugin marketplace add scalably-io/agent-skills` then `/plugin install seo-ops@scalably-agent-skills`, then run `/seo-ops:anchor-policy`.
- Codex or any Agent Skills runtime: `git clone https://github.com/scalably-io/agent-skills && cd agent-skills`, then `cp -R skills/anchor-policy .agents/skills/`. Skills that link to sibling skills expect those siblings to be alongside them: copy the whole `skills/` directory instead of a single folder to keep those links working.
- Raw file: `curl -s https://scalably.io/skills/anchor-policy.md -o SKILL.md` This skill ships references; the raw file alone is incomplete, use the plugin or the folder copy.

## Requirements
- None; input is a draft file and a target URL (plus the anchor spec described below). No external tools or network access needed.

## Inputs and outputs
| | |
|---|---|
| Input | A draft (markdown or plain text) plus an anchor spec: article type (`Standard`/`PR` or `Listicle`), Anchor 1 text + URL, and Anchor 2 text + URL if a second anchor exists |
| Output | A pass/fail JSON verdict: `{"pass": bool, "failed_rules": [...], "notes": [...]}` |

## Reliability
- Release v1.0.0. Structural test suite and `claude plugin validate` run on every push: https://github.com/scalably-io/agent-skills/actions/workflows/validate.yml
- Private client-data scan before every release. Last end-to-end verification: 2026-09-08.
- Derived from production, genericized by hand, reviewed per skill and as a whole before publish.

## Full skill
## What it does

Validates anchor placement, URL correctness, and SEO positioning in a guest-post draft against the canonical anchor rules, for both anchor models a guest post can use: **Standard/PR** (rules S1-S9) and **Listicle** (rules L1-L9). The two models are mutually exclusive; the skill reads the article type first and applies only the matching rule set. Returns a pass/fail verdict with the exact rule(s) violated and actionable remediation.

## Requirements

- None; input is a draft file and a target URL (plus the anchor spec described below). No external tools or network access needed.

## Inputs and outputs

| | |
|---|---|
| Input | A draft (markdown or plain text) plus an anchor spec: article type (`Standard`/`PR` or `Listicle`), Anchor 1 text + URL, and Anchor 2 text + URL if a second anchor exists |
| Output | A pass/fail JSON verdict: `{"pass": bool, "failed_rules": [...], "notes": [...]}` |

## Worked example

Paste into Claude Code with this skill installed:

```text
/seo-ops:anchor-policy check this draft against the anchor spec below.

Spec:
- Article type: Standard
- Anchor 1: "cloud backup service" -> https://example.org/backup

Draft:
## Why Backups Matter
Losing data is costly, so many teams turn to [cloud backup](https://example.org/backup) to stay safe.
```

Expected verdict:

```json
{
  "pass": false,
  "failed_rules": ["S4"],
  "notes": ["Anchor text is \"cloud backup\" but the spec requires the exact text \"cloud backup service\" (case-sensitive). S4 fails. Fix: change the link text to match the spec exactly."]
}
```

## Procedure

### 1. Read the canonical rules

Read `references/canonical-rules.md` in this skill folder. Determine the article type from the spec (Standard/PR or Listicle) before checking anything else; the two rule sets contradict each other on purpose and must never be mixed.

### 2. Validate

Apply every rule in the matching set (S1-S9 for Standard/PR, L1-L9 for Listicle) using the "For QA (validation phase)" section of `canonical-rules.md`. A missing anchor, a URL mismatch, or an unauthorized/competitor link fails QA immediately regardless of any other rule's status.

### 3. Report

Return JSON:

```json
{
  "pass": true,
  "failed_rules": [],
  "notes": []
}
```

When failing, set `pass=false`, list every violated rule ID in `failed_rules` (e.g. `"S2"`, `"L6"`), and give one concrete, actionable remediation per failure in `notes`.

<verification>
Before returning a verdict:
1. Every rule in the matching set (S1-S9 or L1-L9) was checked explicitly; no silent omissions.
2. Each failed rule has a specific reason tied to the draft text, not a generic statement.
3. Remediation guidance says exactly what to change, not just what is wrong.
4. Never mark a rule compliant when the evidence needed to check it (draft text, anchor spec) is missing; fail safe (`pass=false`) instead, and say what evidence is missing.
</verification>