seo-agent 0.4.0

Lightweight CLI that runs a technical SEO audit for a URL and outputs prioritized, actionable recommendations.

SEO Audit Agent

Dependency-free technical SEO audit CLI for quick, actionable site reviews. Built for developers, SEO engineers, and teams who want fast, deterministic audits that work locally or in CI.

Why SEO Audit Agent

• Dependency-free runtime (stdlib only) with fast setup
• Actionable recommendations prioritized by impact, effort, and confidence
• Optional crawl sampling and link checks to catch template-level issues
• CI-friendly JSON/Markdown output, baseline diffs, and quiet mode
• Offline enrichers for PageSpeed and Search Console exports

Quick start

seo-agent https://example.com --goal "traffic growth"

If running from source:

python3 -m seo_agent https://example.com --goal "traffic growth"

• If --goal is omitted, the agent asks for your main objective before auditing.
• If you hit SSL certificate errors, re-run with --insecure (only when you trust the site).

Requirements

• Python 3.9 or newer
• Network access to fetch the target page and robots.txt

Installation

PyPI (recommended)

pipx install seo-agent
# or
pip install seo-agent

If the package is not published yet, install from source:

From source

git clone https://github.com/ShubhenduVaid/seo-agent.git
cd seo-agent
python3 -m venv .venv
source .venv/bin/activate
python3 -m pip install --upgrade pip  # no additional packages required
python3 -m pip install -e .

Usage

seo-agent <url> [--goal "primary objective"] [--insecure]

Examples:

• seo-agent https://example.com --goal "traffic growth"
• seo-agent https://example.com --insecure
• seo-agent https://example.com --format json --quiet (machine-readable output)
• seo-agent https://example.com --fail-on-critical (exit non-zero if critical issues found; good for CI)
• seo-agent https://example.com --crawl-depth 1 --crawl-limit 5 (sample a handful of internal pages)
• seo-agent https://example.com --crawl-sitemaps --crawl-limit 8 (seed crawl from sitemaps)
• seo-agent https://example.com --crawl-depth 1 --crawl-delay 0.5 (polite crawl with delay; honors robots.txt crawl-delay)
• seo-agent https://example.com --crawl-include "/blog/*" --crawl-exclude "*/tag/*" (scope crawl sampling)
• seo-agent https://example.com --report /tmp/report.txt (also write the report to a file)
• seo-agent --list-checks (show available checks)
• seo-agent --version
• seo-agent --config ./seo-agent.ini https://example.com (use shared defaults from a config file)
• seo-agent https://example.com --format sarif --report ./reports/seo.sarif (SARIF for code scanning)
• seo-agent https://example.com --format github (GitHub Actions summary format)

For backward compatibility you can also run python3 seo_agent.py ... from the project root.

The report is grouped by severity:

1. Critical Issues - fix immediately (high impact)
2. Important Optimizations - fix soon (medium impact)
3. Recommended Enhancements - nice to have

Each issue includes what is wrong, why it matters, step-by-step fixes, expected outcome, and how to validate.

• Reports include HTTP status, a simple score, and top 5 priorities. JSON output includes scores.
• Response time and document size are included for quick Web Vitals triage.
• Goal-aware scoring slightly boosts performance/content/linking issues when goals mention traffic/growth.
• Crawl summary highlights duplicate titles/descriptions across sampled pages.

Crawl filters use glob patterns against URLs or paths (e.g., /blog/*, */search*). Excludes always win.

What it checks

• Site speed signals: page weight, script count, render-blocking scripts, resource hints, image sizing, lazy-loading hints (LCP/FID/CLS risk proxies)
• Static asset hygiene: cache-control and compression hints for sampled JS/CSS via HEAD requests
• Crawlability: robots.txt availability/content, sitemap discovery, meta robots directives, X-Robots-Tag
• Polite crawling: optional limited crawl that honors robots.txt disallow/crawl-delay and rate limits requests
• Redirects: detects when the requested URL redirects to a different host/path
• Mobile optimization: viewport tag and lazy-loading coverage
• Security: HTTPS presence and HSTS header hint
• Security headers: Content-Security-Policy, Referrer-Policy, X-Content-Type-Options, Permissions-Policy, X-Frame-Options
• Response health: HTTP status reporting (4xx/5xx) for the audited URL
• Structured data: JSON-LD detection
• Internal linking: ratio of internal/external links, low internal link coverage
• Duplicate control: canonical tag presence, host consistency, follow directives
• Meta and headings: title quality, description presence and length, social meta completeness, H1 usage, hreflang x-default hint and absolute hrefs, image alt coverage

Sample output (truncated)

Primary goal: traffic growth
URL audited: https://example.com

1. Critical Issues - fix immediately (high impact)
- Title tag missing
  What: No <title> found; search results will lack a meaningful headline and relevance signal.
  Fix steps:
    - Add a concise, descriptive <title> (50-60 chars) targeting the primary keyword.
    - Place the most important terms first and keep branding at the end.
    - Avoid duplicating titles across pages; keep them unique.
  Outcome: Stronger relevance signals and improved CTR from SERPs.
  Validate: View source to confirm the title; check Search Console HTML improvements for duplicates.

Output formats

• Default text
• --format json for structured output (good for CI)
• --format markdown for docs/issue comments
• --format sarif for GitHub code scanning / SARIF-compatible tooling
• --format github for GitHub Actions job summaries
• --report <path> to write the rendered output to a file
• --quiet skips interactive prompts (useful in CI)
• --fail-on-critical exits non-zero if critical issues are found (useful for CI gates)

Configuration file

Use an INI file to keep repeatable defaults and team presets.

[seo-agent]
url = https://example.com
goal = traffic growth
format = markdown
crawl_depth = 1
crawl_limit = 8
crawl_delay = 0.5
crawl_sitemaps = true
crawl_include = /blog/*
crawl_exclude = /search*, /tag/*
check_links = true
report = reports/seo-report.md
fail_on_critical = true

Run with:

seo-agent --config ./seo-agent.ini

CLI flags always override config values. Unknown keys in the config will be reported unless --quiet is set.

GitHub Action

Use the official GitHub Action to run audits in CI and publish a job summary:

- uses: ShubhenduVaid/seo-agent@v0.4.0
  with:
    url: https://example.com
    goal: traffic growth
    format: github
    fail_on_critical: true
    extra_args: --crawl-depth 1 --crawl-limit 5 --crawl-exclude "/search*"

See docs/GITHUB_ACTION.md for full usage.

Baselines and comparisons

• --save-baseline <path> saves a JSON snapshot of issues
• --compare <path> compares current issues to a previous baseline
• Useful for tracking improvements across releases and migrations

Integrations (offline)

• --psi-json <path>: include PageSpeed/Lighthouse metrics from a local JSON export
• --gsc-pages-csv <path>: weight priorities using Search Console Pages export data

See docs/INTEGRATIONS.md for instructions on generating these files.

Documentation

• docs/CLI_REFERENCE.md - full CLI reference and examples
• docs/ARCHITECTURE.md - current architecture and data flow
• docs/OUTPUT_SCHEMA.md - JSON output schema for --format json
• docs/INTEGRATIONS.md - PageSpeed and Search Console offline enrichers
• docs/TROUBLESHOOTING.md - common issues and fixes
• docs/ROADMAP.md - project direction and future features
• docs/DISCOVERABILITY.md - GitHub visibility checklist
• docs/GITHUB_ACTION.md - GitHub Action usage
• CHANGELOG.md - release notes

Development

Run the CLI locally while iterating:

seo-agent https://example.com --goal "traffic growth"

Run tests:

python3 -m unittest discover -v

Lint and type check (optional):

python3 -m pip install -r requirements-dev.txt
python3 -m ruff check .
python3 -m mypy seo_agent

Project layout (key modules):

• seo_agent/cli.py - CLI argument parsing and entry point
• seo_agent/audit.py - audit orchestration + crawl sampling
• seo_agent/analyzer.py - HTML parser used by audits
• seo_agent/checks/ - built-in checks + registry (optional plugins)
• seo_agent/network.py - network helpers (fetching, robots, normalization)
• seo_agent/robots.py - robots.txt parsing and allow/disallow matching
• seo_agent/baseline.py - baseline save/compare (diffs)
• seo_agent/integrations/ - optional offline data enrichers (PageSpeed/GSC exports)
• seo_agent/reporting.py - report rendering and formatting
• tests/ - unit tests for core utilities and checks

Packaging and release

Build a wheel/sdist locally (requires build if not already installed):

python3 -m pip install --upgrade build
python3 -m build

This produces artifacts under dist/. Update the version in seo_agent/__init__.py and pyproject.toml, and add notes to CHANGELOG.md before tagging a release.

GitHub Actions CI:

• Pull requests and main branch: installs in editable mode and runs lint (ruff), mypy, and python -m unittest discover -v with coverage >= 70%.
• Tag pushes matching v*: builds sdist/wheel and publishes to PyPI using OIDC (pypa/gh-action-pypi-publish). Configure PyPI trusted publisher for the repo before tagging.

Contributing

Contributions are welcome! Please read CONTRIBUTING.md for filing issues, proposing features, and submitting pull requests.

Security

To report a vulnerability, follow the process outlined in SECURITY.md. Please avoid filing public GitHub issues for security reports.

License

This project is available under the MIT License. See LICENSE for details.