What does this skill do, and when should you use it?
geo-prospect is the business-side skill in the geo-seo-claude repository (a bundle of 16 skills), acting as a lightweight CRM. It stores prospect records as a local JSON file (~/.geo-prospects/prospects.) and manages a five-stage pipeline: Lead → Qualified → Proposal Sent → Won → Lost. Each record can carry audit snapshots, timestamped interaction notes, monthly contract value, and loss reasons, with a visual pipeline summary including revenue forecasting. It integrates with the sibling /geo quick audit command to write GEO scores directly into prospect records. Data persists across sessions and survives uninstallation.
Reads and writes prospect records in ~/.geo-prospects/prospects., each with a sequential ID (PRO-001 onward), company name, domain, contacts, status, GEO score, deal value, and notes. Offers ~10 slash commands: /geo prospect new (interactive creation, auto-detects company name from domain), list (filterable by status), show, audit (runs /geo quick and archives the score), note (adds timestamped remarks), status (moves pipeline stage), won/lost (records contract value or loss reason), and pipeline (outputs stage counts, committed MRR, total pipeline value, and suggested next actions). Audit snapshots and proposals are saved to audits/ and proposals/ subdirectories.
- Solo GEO consultant: batch-enter prospects by domain, run quick audits, and route anyone scoring below 55 straight to proposal generation.
- Small agency sales lead: run /geo prospect pipeline weekly to review committed MRR, pipeline value, and follow-ups due (e.g., a proposal sent 8 days ago).
- Client handover: use show to present a client's full audit history and interaction notes so a colleague can pick up the account.
- Loss retrospective: log why deals died (e.g., budget freeze) with the lost command for future reference.
- Ops/back-office: manage client data as plain local JSON instead of a SaaS CRM, keeping everything on your own machine.
How do you install this skill?
- Prompt-only skill with no script implementation or test coverage; actual behavior cannot be verified statically.
- No backup/rollback/corruption-recovery mechanism for prospects.; manual JSON edits risk data loss — back up ~/.geo-prospects/ before use.
- Example data (companies, dates, revenue forecasts) is fabricated demonstration content, not evidence of real capability.
- The audit command depends on sibling skills (/geo quick); installing this skill alone leaves functionality incomplete.
- No Chinese-language support; documentation targets English/Italian market users.
- Shell / CLI
- Local filesystem
Claude Code CLI (for the parent geo-seo-claude collection, including /geo quick audits invoked by this skill)
Install the parent repo geo-seo-claude (this skill ships inside it): on macOS/Linux run curl -fsSL https://raw.githubusercontent.com/zubair-trabzada/geo-seo-claude/main/install.sh | bash; on Windows run install-win.sh from Git Bash. Requires Python 3.8+, Claude Code CLI, and Git (optional: uv for faster installs, Playwright for screenshots). The skill lives at skills/geo-prospect/SKILL.md; Python dependencies install into an isolated venv at ~/.claude/skills/geo/.venv/.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/zubair-trabzada/geo-seo-claude.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/geo-prospect" ~/.claude/skills/
rm -rf "$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?
Once installed, send your agent any of these to trigger it:
- asked for quote
In Claude Code, say 'prospect', 'lead', 'client', or 'pipeline', or use commands directly: /geo prospect new electron-srl.com creates a prospect; /geo prospect audit electron-srl.com scores and archives it; /geo prospect note PRO-001 "asked for quote" adds a note; /geo prospect status PRO-001 qualified advances the stage; /geo prospect won PRO-001 4500 records monthly contract value; /geo prospect pipeline shows the revenue-forecast summary. If ~/.geo-prospects/ doesn't exist, create it first with mkdir -p ~/.geo-prospects/audits ~/.geo-prospects/proposals (the skill says it also creates directories automatically). Note: the audit command depends on the sibling /geo quick skill and cannot score without the rest of the collection installed.
What are this skill's strengths and limitations?
- Data is plain local JSON — clear structure, easy to back up or migrate, no third-party SaaS dependency.
- Well-defined pipeline stages (lead/qualified/proposal/won/lost) with automatic trail-keeping: audit scores, timestamped notes, loss reasons.
- Tight loop with /geo quick audits: GEO scores land directly in the prospect record, and scores under 55 trigger a proposal suggestion.
- The pipeline command outputs committed MRR, pipeline value, and concrete next actions — useful for weekly reviews.
- Hardcoded to the GEO agency workflow — not a general-purpose CRM; no custom fields or custom stages.
- Audit features depend on sibling geo-audit skills and the installed Python environment; not usable standalone for scoring.
- Single-user, single-file JSON database: no concurrency protection, no multi-user permissions, questionable at scale.
- No test suite; README market statistics (4.4x conversion, $850M market) are marketing citations without verifiable sources.
- Pipeline summaries show euro-denominated sample figures; currency handling and exchange rates are unspecified.
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 Prospect Manager (geo-prospect) this page | 45 · Not recommended | ★ 11k | 20d ago | MIT |
| GEO Monthly Delta Report Generator | 46 · Use with care | ★ 11k | 20d ago | MIT |
| GEO Proposal Generator | 44 · Not recommended | ★ 11k | 20d ago | MIT |
| Linear Claude Skill | 55 · Use with care | ★ 135 | 3mo ago | MIT |
| Kanban AI | 52 · Use with care | ★ 25 | 7mo ago | Apache-2.0 |
How did FollowSkills review this skill?
Data flow is transparent (all writes to ~/.geo-prospects/, README discloses the uninstaller does not remove this directory), no network exfiltration or credential access, declared tools Read/Write/Bash/Glob. But Bash scope is unconstrained, JSON writes lack user confirmation steps, and there is no backup, rollback, or corruption-recovery path — deducted accordingly.
Instructions are self-consistent, the command table is complete, and storage structure is clear, so the happy path is plausible; however, no tests cover this skill (repo tests only cover fetch_page.py SSR detection, unrelated), and there are no provisions for concurrent writes, JSON corruption, missing fields, or defined failure feedback. Scored below the static cap.
Target scenario (GEO agency pipeline management) is clear, trigger words are listed, and pipeline stages are well defined; but capability boundaries and non-fit ranges are undeclared, the audit command depends on the sibling /geo quick skill without stating that coupling here, there is no Chinese-language support, and generic trigger words like 'lead'/'client' raise false-trigger risk.
Frontmatter includes version, tags, and allowed-tools; docs are well layered (command table, data structure, pipeline definitions, storage layout) with MIT license at repo level; but no changelog, no declared maintenance ownership or update path, example data (Electron Srl, 2026-03-12) appears to be fabricated demo content rather than real records, and known limitations are not disclosed.
The core task (JSON-record CRUD and pipeline summaries) is logically complete with defined output formats, and the marginal value is structured persistence; but it is purely prompt instructions whose execution quality depends entirely on the model, with no verified output samples — the revenue-forecast figures in examples are demonstrative only.
Only author-written documentation and fabricated example outputs exist; no tests target this skill, no reproducible evidence, and example data (dates, companies, scores) cannot be cross-corroborated; promotional content (Skool community funnel) is mixed into repository context, hence a low score.
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 →