Rank Tracker
Monitor keyword positions, SERP features, and AI-search visibility over time.
The skill limits WebFetch to optional retrieval, requires source labels for measurements, and asks before saving results. Points are deducted because sensitive-data handling, data minimization, rollback for memory writes, and explicit authorization for external fetching are not defined in the skill itself; fetched-content safety is mainly documented at repository level.
The method, completion criteria, missing-baseline behavior, N/A handling, and stop conditions are reasonably coherent, with ledger commands and some failure guidance. Static evidence does not establish that the referenced scripts, connectors, or host integrations run successfully, and several paths and tool placeholders depend on the host environment, so the score is capped below 10.
Triggers, audience scenarios, inputs, output forms, and the non-fit boundary for stakeholder reports and alerts are explicit, with bilingual naming and a Chinese trigger. Points are deducted for limited evidence of host compatibility, mainland-China network reachability, Chinese SERP behavior, and optional-tool coverage for competitor, GEO, and SERP data.
The skill has frontmatter, versioning, Apache-2.0 licensing, Quick Start prompts, a shared contract, linked templates, reference guidance, and a next-skill handoff; repository context adds shared CI and security documentation. Points are deducted because the skill itself does not state a maintainer, changelog/update path, installation procedure, troubleshooting guidance, or a complete worked output.
It covers snapshots, movement analysis, SERP features, AI visibility, competitor comparison, and source-tagged reporting, while explicitly prohibiting invented data; the templates are close to directly usable. Points are deducted because static files do not verify real retrieval accuracy or low-review usability, and causal explanations are explicitly hypotheses rather than confirmed diagnoses.
Metric provenance labels, baseline requirements, ledger snapshots, and source-tagged templates provide some auditability. Points are deducted because there are no real execution records, independent corroboration, or dedicated key-path tests for this skill; the repository tests concern architecture and artifact validation rather than rank-tracking reproduction.
- WebFetch and optional SEO/AI tools may involve third-party data transfer, regional reachability, and service-specific behavior; confirm authorization, privacy boundaries, and robots/TOS requirements first.
- Do not present Estimated values or causal hypotheses as measured conclusions; mark unavailable data N/A and review it manually.
- The skill describes writes to memory/open-loops.md and memory/monitoring/, but does not define write permissions, failure recovery, or rollback behavior in this file.
What does this skill do, and when should you use it?
Rank Tracker compares keyword rankings and SERP performance across time. It reads tool exports, user-supplied data, or existing baselines and produces source-labeled ranking reports or delta summaries. Beyond traditional positions, it can track featured snippets, PAA, image or video packs, local results, AI Overview presence, and citation visibility. It fits SEO/GEO teams that need traceable ranking monitoring, but it is not intended for multi-metric reporting or alert configuration.
Reads current rankings, prior baselines, target keywords, market and device data, plus user-provided or tool-derived metrics; can use WebFetch and the repository's Python ledger commands to record ranking snapshots, calculate deltas, and inspect trends; compares SERP features, AI visibility, and competitors; outputs current position, baseline, labeled movement, source, ranking URL, feature ownership, major gains and losses, likely causes, and follow-up recommendations; after confirmation, results can be saved to memory/monitoring/.
- An SEO lead has several ranking exports and needs to compare the latest snapshot with the previous one.
- A content team wants to identify which target keywords reached page one and which declined materially.
- A GEO team needs to record AI Overview presence, citation rate, or citation position when some data is not connected.
- An agency needs a client-ready ranking summary with Measured, User-provided, and Estimated source labels.
- A team without a paid SEO platform wants to build a snapshot-based ranking trend using supplied data or the ledger.
What are this skill's strengths and limitations?
- Requires every ranking and metric to carry a source label: Measured, User-provided, or Estimated.
- Covers baseline deltas, trends, SERP features, competitors, and AI visibility.
- Supports snapshot-based measurement without paid tools and forbids narrating unmeasured movement.
- Uses the repository's shared contract and produces a monitoring handoff summary.
- A keyword list is required unless keywords can be inferred from existing context.
- SERP features, AI citations, and competitor data depend on supplied data or available tools and may be N/A.
- Ranking changes are week-scale, confounded outcomes; the skill cannot establish a single causal explanation by itself.
- The source provides no standalone test-suite evidence, platform-coverage testing, or detailed configuration for specific third-party SEO tools.
How do you install this skill?
Install the full repository with: npx skills add aaron-he-zhu/aaron-marketing-skills. You can also clone it with: git clone https://github.com/aaron-he-zhu/aaron-marketing-skills, then use seo-geo/evaluate/rank-tracker/SKILL.md. In Claude Code, run /plugin marketplace add aaron-he-zhu/aaron-marketing-skills followed by /plugin install aaron-marketing@aaron. The source does not document a dedicated command for copying only this skill into another host.
How do you use this skill?
Provide a domain and keyword list, for example: Set up rank tracking for example.com targeting these keywords: [keyword list]. You can also ask: Analyze ranking changes for example.com over the past 30 days. If no keywords are supplied or inferable from CLAUDE.md or prior monitoring records, the skill asks whether to provide a list, track the domain's known top terms, or cancel. Without a baseline, it records the current run as the first snapshot; missing optional SERP or AI-citation data is marked N/A. It asks whether to save the results when finished.
How does this skill compare with similar options?
Unlike performance-monitor, Rank Tracker focuses on changes in keyword positions, SERP features, and AI visibility; performance-monitor is the repository's option for multi-metric stakeholder reports or alert setup.