What does this skill do, and when should you use it?
seo-flow is one skill within the Claude SEO repository, integrating 41 FLOW AI prompts across Find, Leverage, Optimize, Win, and Local stages. It selects and applies prompts based on the requested stage, URL or topic, and—during optimization—industry, prior analysis, and URL signals. Its scope includes keyword research, gap analysis, SERP intent mapping, backlink strategy, content optimization, conversion, and local SEO. It suits practitioners who want a staged SEO prompting workflow, but it depends on Claude Code’s /seo command behavior and is not the repository’s full audit suite.
Reads the FLOW framework reference on every /seo flow activation and loads prompt files only for the requested stage; applies Find prompts for keyword research, gap analysis, and SERP intent mapping; applies Leverage prompts for backlink and off-site authority strategy; selects exactly 2–3 Optimize prompts using industry, previous analysis, and URL signals; applies Win prompts for BOFU, conversion-rate, and dual-surface scorecards; applies Local prompts for GBP, metadata, title tags, and local audits; displays the complete 41-prompt index; and runs scripts/sync_flow.py to synchronize prompt files and display a JSON summary.
- An SEO consultant starting work on a URL or topic can use `/seo flow find` for keyword research, content-gap analysis, and SERP-intent mapping.
- An SEO team with existing audit findings can use `/seo flow optimize` to select prompts relevant to the industry, page type, and previously identified issues.
- A marketer reviewing a conversion-focused page can use `/seo flow win` for BOFU and conversion-oriented analysis.
- A local business or local SEO consultant can use `/seo flow local` for GBP, metadata, title-tag, and local-audit prompts.
- A practitioner can use `/seo flow prompts` to inspect all prompts and trigger conditions, or `/seo flow sync` to update the prompt files.
How do you install this skill?
- Before running `/seo flow sync`, verify the remote source, pin a revision, and retain a recoverable local copy; otherwise the prompt library may change without confirmation.
- Treat framework statistics and SEO recommendations as claims requiring review, not as verified conclusions; check source, date, geography, and current validity before publication.
- Users in mainland China should verify reachability of GitHub and related external resources and prepare local references or a mirror.
- Shell / CLI
- Network access
- Local filesystem
Python 3.10+
Install the complete Claude SEO collection as documented in the README: in Claude Code, run /plugin marketplace add AgriciDaniel/claude-seo, then /plugin install claude-seo@agricidaniel-claude-seo. Alternatively, clone the repository and run bash claude-seo/install.sh. The source does not document a separate installation procedure for seo-flow in other clients.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/AgriciDaniel/claude-seo.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/seo-flow" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
In Claude Code, run /seo flow for the stage menu. Use /seo flow find [url|topic], /seo flow leverage [url], /seo flow optimize [url], /seo flow win [url], or /seo flow local [url] for stage-specific work. Use /seo flow prompts to list all 41 prompts and /seo flow sync to synchronize them. Every activation outputs the FLOW attribution notice under CC BY 4.0.
What are this skill's strengths and limitations?
- Organizes 41 SEO prompts by workflow stage instead of relying on one generic prompt.
- Uses industry, prior analysis, and URL signals to select 2–3 relevant Optimize prompts.
- Covers research, SERP intent, backlink strategy, content optimization, conversion, and local SEO use cases.
- The prompt library is attributed under CC BY 4.0, while the skill metadata specifies MIT licensing.
- It relies on Claude Code’s /seo flow slash-command behavior and reference-file layout, so Codex or other clients require adaptation.
- The source provides no separate test results or standalone installation package for seo-flow.
- The Leverage stage contains only one prompt, giving it narrower coverage than the other stages.
- Sync requires a Python script and network access; missing reference files prevent normal analysis.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| FLOW Evidence-Led SEO this page | 47 · Use with care | ★ 19k | 23d ago | MIT |
| Blog Cluster Engine | 53 · Use with care | ★ 2.4k | 13d ago | MIT |
| DataForSEO Live SEO Intelligence | 47 · Use with care | ★ 19k | 23d ago | MIT |
| SERP Analysis | 56 · Use with care | ★ 3k | 13d ago | Apache-2.0 |
| SEO Content Brief Generator | 52 · Use with care | ★ 19k | 23d ago | MIT |
How did FollowSkills review this skill?
The skill requires evidence use, attribution, and explicit missing-file messages; the MIT license and repository security material add some safety context. Points are deducted because `/seo flow sync` runs a script that pulls content from GitHub without requiring confirmation, version verification, or rollback, while data-flow, external-request boundaries, and sensitive-input handling remain underspecified.
Commands, stage routing, on-demand loading, and several missing-file responses are clearly described, making the happy path plausible. Points are deducted because key paths were not executed in this static review; several prompt files are substantially duplicated, inputs for stages such as `leverage` are underspecified, and diagnosis of network, authentication, and abnormal-input failures is delegated to external tooling. The static ceiling keeps this below 10.
The audience and trigger phrases are reasonably explicit, with URL-or-topic inputs and find, leverage, optimize, win, and local scenarios. Points are deducted because non-fit boundaries, input-quality requirements, and output-format limits are not clear; Chinese-language interaction is not declared, and core synchronization depends on GitHub and other overseas services whose mainland-China reachability is uncertain.
The SKILL.md provides a command table, orchestration logic, reference-file index, attribution requirements, and error handling; version, author, MIT licensing, and CC BY 4.0 source information are present. Points are deducted because maintenance ownership, changelog/update responsibility, dependency installation, sync security, version locking, and troubleshooting are incomplete within the selected skill itself.
The staged prompts, input lists, output lists, evidence requirements, and measurement planning can produce directly usable SEO working documents and provide some value beyond manual prompting. Points are deducted because the skill mainly orchestrates prompts, output quality depends on the model, supplied evidence, and other reference files, and no representative FLOW execution result is statically verifiable. The static ceiling keeps this at or below 7.
The reference files include sources, retrieval dates, licenses, hashes, and claim citations, and the framework requires separating evidence from assumptions; repository CI and tests provide additional context but do not cover the selected FLOW paths. Points are deducted because actual execution, output correctness, and cross-source corroboration cannot be confirmed, so the static ceiling limits this to 5.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →