What does this skill do, and when should you use it?
geo-compare is one of the 16 skills bundled in the geo-seo-claude repository, focused specifically on monthly progress tracking for GEO (Generative Engine Optimization) clients. It reads a baseline audit and a current audit, computes deltas across the overall GEO score, 6 categories, 5 AI platforms, and 14 AI crawler access states, and generates a complete Markdown client report with charts, action-item status, monthly wins, and next-month priorities. Reports are saved under ~/.geo-prospects/reports/. It does not audit sites itself — without prior audit files it will point you to /geo prospect audit first. It fits GEO agencies that already use the repo's other audit skills and need to demonstrate measurable value to clients each month.
Triggered via /geo compare <domain> or by passing two audit file paths directly. Workflow: 1) finds audit files for the domain in ~/.geo-prospects/audits/, sorts by date, uses the oldest as baseline and newest as current (if only one exists, it serves as baseline and a fresh quick audit is run as current); 2) parses both audits for overall score, 6 category scores, 5 platform scores, allow/block status of 14 AI crawlers, critical issues, and action items; 3) calculates deltas with trend symbols (▲ improved, ▼ declined, ── unchanged); 4) generates a Markdown monthly report (ASCII score bars, before/after comparison tables, action-plan completion status, monthly wins, newly discovered issues, next-month priorities, 6-month trajectory, and estimated business impact) saved to ~/.geo-prospects/reports/<domain>-monthly-<YYYY-MM>.md, plus a printed summary. When exact scores are absent from audit files, it estimates them from the written findings.
- A GEO agency generates a monthly progress report before a client check-in with /geo compare electron-srl.com --month march-2026, showing the score moved from 32 to 44
- An account manager uses the action-item completion table (e.g., 3/5 quick wins done) as renewal-negotiation evidence of delivered work
- With only one historical audit on file, the skill uses it as baseline and runs a fresh quick audit to produce a first delta report immediately
- An agency lead reviews the 6-month trajectory table to confirm a client is on track toward the Month-6 target of 65/100
- When a score declined, the 'New Issues Discovered' section helps reframe the drop as newly found optimization opportunities
How do you install this skill?
- Static review only — no code was executed; scores are based on source reading with low confidence.
- The 'estimate scores from context' mode can produce monthly reports containing numbers without factual grounding; manually verify before client delivery.
- The report template's 'Estimated Business Impact' placeholders (AI traffic %, euro value) invite inflated or fabricated marketing figures.
- The skill depends on a specific Markdown audit format produced by the same repo's geo-prospect audit; compatibility with other formats is undeclared.
- Bash is an allowed tool with no stated confirmation or rollback mechanism; users should review executed commands.
- Publisher identity is unverified; maintenance and update path unclear; no Chinese-language support.
- Shell / CLI
- Local filesystem
Claude Code CLI (slash commands /geo compare)existing geo-seo-claude skill collectionaudit files in ~/.geo-prospects/audits/
The skill cannot be installed standalone; install the whole geo-seo-claude repository. macOS/Linux: curl -fsSL https://raw.githubusercontent.com/zubair-trabzada/geo-seo-claude/main/install.sh | bash. Windows (Git Bash): use install-win.sh. Manual: git clone https://github.com/zubair-trabzada/geo-seo-claude.git && cd geo-seo-claude && ./install.sh. Requirements: Python 3.8+ (plus python3-venv on Debian/Ubuntu), Claude Code CLI, Git; optional uv for faster installs and Playwright for screenshots. MIT licensed.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/zubair-trabzada/geo-seo-claude.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/geo-compare" ~/.claude/skills/
rm -rf "$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?
After installing, open Claude Code and run: /geo compare <domain> (e.g., /geo compare electron-srl.com); or specify two audit files directly: /geo compare ~/.geo-prospects/audits/<domain>-2026-01-15.md ~/.geo-prospects/audits/<domain>-2026-03-12.md; optionally with a month: /geo compare electron-srl.com --month march-2026. If no audit files exist, it will suggest running /geo prospect audit <domain> first. Afterwards, you can run /geo report-pdf for a visual version (that PDF capability belongs to a sibling skill in the same repo).
What are this skill's strengths and limitations?
- Complete, retention-oriented report structure: score progress bars, category/platform before-after tables, crawler access changes, action-item completion rates, and a 6-month trajectory
- Accepts both a domain and explicit file paths, with clearly documented file-discovery logic (oldest = baseline, newest = current)
- Includes trend-interpretation rules, including talking points to reframe score declines as newly discovered opportunities
- Built-in trigger phrases (compare, delta, monthly report, plus Italian equivalents) make model-invocation reliable
- Depends heavily on sibling skills (/geo prospect audit, /geo report-pdf) — standalone it has no data source and is nearly unusable
- When audit files lack exact scores, it estimates scores from written findings, which weakens the reliability of the delta comparison
- Business-impact figures in the report (AI citation lift %, monthly traffic value in €) are template placeholders, not measured data
- No repo topics/tags and no test suite or sample report output for this sub-skill
- Monthly output is Markdown only; the visual PDF requires another skill from the repo
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 |
|---|---|---|---|---|
| GEO Monthly Delta Report Generator this page | 46 · Use with care | ★ 11k | 20d ago | MIT |
| GEO Proposal Generator | 44 · Not recommended | ★ 11k | 20d ago | MIT |
| GEO PDF Report Generator | 50 · Use with care | ★ 11k | 20d ago | MIT |
| GEO Prospect Manager (geo-prospect) | 45 · Not recommended | ★ 11k | 20d ago | MIT |
| AI Legal Assistant (legal/SKILL.md Orchestrator) | 43 · Not recommended | ★ 1.8k | 7mo ago | — |
How did FollowSkills review this skill?
Skill only reads local audit files and writes to ~/.geo-prospects/reports/; no network calls or sensitive-data flows, storage location disclosed in README, and uninstaller preserves user data (transparent). Deducted for: Bash allowed without stated confirmation mechanisms, no rollback guidance, and incomplete data-flow disclosure.
Workflow is self-consistent: file discovery, parsing, delta calculation and output paths are specified, with a helpful failure path when no audits exist. However, parsing relies on brittle text-pattern matching, and the skill explicitly permits 'estimating scores from context' when markers are missing — introducing subjectivity and irreproducibility. Repo tests cover only fetch_page.py SSR detection, not this skill's key paths. Static cap 10; deducted for non-deterministic estimation mode and untested key paths.
Triggers are explicit (English and Italian), the target scenario (GEO agency monthly client reporting) is clear, and inputs/outputs/preconditions are declared. Deducted for: narrow boundary depending on Markdown audit files in ~/.geo-prospects/audits/ with no guidance for other input sources, no Chinese support, and unverified mainland-China reachability (limited impact since no external service dependency).
SKILL.md is versioned with frontmatter and well-layered structure (purpose/commands/workflow/template/calculation logic); README documents architecture and data storage; MIT license present. Deducted for: no changelog, no known-limitations section, unclear maintenance ownership (unverified publisher), and marketing content embedded in the template.
Core task (comparing two audits and generating a progress report) is fully specified with a detailed, directly generatable markdown template. However, the estimation mode can produce baseline/current figures without factual grounding, and the 'Estimated Business Impact' section invites fabricated percentages/euro values, so output requires human verification before client delivery. Static cap 7; deducted for limited direct usability and thin comparative-benefit evidence.
Example figures (32→44) are demonstrative, not verified outputs; delta logic is only a pattern list; repo tests do not cover this skill; no third-party execution evidence. Static cap 5; deducted for absence of auditable reproduction material for this skill's key paths.
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 →