Rank Tracker
Tracks keyword positions, SERP feature ownership, and AI visibility over time.
Quick Start
Set up rank tracking for [domain] targeting these keywords: [keyword list]
Analyze ranking changes for [domain] over the past [time period]
Skill Contract
Expected output: a ranking report or delta summary plus the standard handoff summary for memory/monitoring/.
- Reads: current rankings, prior baselines, target keyword list, market/device, and any user-provided or tool metrics.
- Writes: a user-facing monitoring deliverable and reusable summary.
- Promotes: significant changes, confirmed anomalies, follow-up actions, and pending decisions to
memory/open-loops.md.
- Done when: every tracked keyword shows current position vs baseline with a labeled delta (or N/A); each position cites its source (tool export / user-provided / estimated); and biggest movers and likely causes are named.
- Primary next skill: alert-manager when recurring monitoring should become automated.
Handoff Summary
Emit the standard shape from skill-contract.md §Handoff Summary Format.
Data Sources
All integrations optional (see CONNECTORS.md). With tools, pull rankings from ~~SEO tool, impressions from ~~search console, traffic from ~~analytics, and AI citations from ~~AI monitor. Without tools, ask for positions, volumes, competitor data, and SERP feature status.
Zero-dependency measurement loop (no paid tool needed): never narrate a ranking movement you did not measure — record each check as a snapshot and let the ledger compute the delta. python3 "${CLAUDE_PLUGIN_ROOT}/scripts/connectors/ledger.py" record <domain> --source rank --data '{"<keyword>": <position>, ...}', then ledger.py diff <domain> --source rank for the change since the last check and ledger.py trend <domain> --source rank --field "<keyword>" for the series. Record real Search Console positions when connected, or the positions the user provides at Tier 1. A ranking delta is a week-scale, confounded outcome — attribute it against a control (a holdout keyword/competitor), not against zero; see references/measurement-protocol.md. See scripts/connectors/README.md.
Decision Gates
Stop and ask the user when:
- No target keywords are provided and none can be inferred from
CLAUDE.md or prior monitoring records — offer: (1) supply a keyword list, (2) track the domain's top known terms, (3) cancel.
Continue silently (never stop for):
- No prior baseline exists — record the current run as the baseline, label all positions as the first snapshot, and proceed (do not invent a "previous" position).
- Missing optional tool data (SERP features, AI citations) — mark N/A and proceed.
Instructions
When a user requests rank tracking or analysis:
- Set Up Keyword Tracking — configure domain, market, device, language, update frequency, priorities, and competitor watchlist.
- Record Current Rankings — output a position table where every row cites its source (tool export / user-provided / estimated), with position ranges, ranking URLs, feature ownership, and movement vs baseline.
- Analyze Ranking Changes — highlight biggest wins, declines, stable terms, new rankings, lost rankings, likely causes, and recovery ideas; each delta labeled against its baseline.
- Track SERP Features — compare ownership of snippets, PAA, image/video packs, local packs, and related feature shifts.
- Track GEO / AI Visibility — monitor AI Overview presence, citation rate, citation position, and trend; mark each value Measured (from an ~~AI monitor) or N/A if unobserved.
- Compare Against Competitors — report share of voice, head-to-head comparisons, and threat levels.
- Generate Ranking Report — output overall trend, key wins, concerns, opportunities, SERP feature changes, GEO visibility, and recommendations, with each metric carrying its source tag.
Label every metric Measured (tool/export), User-provided, or Estimated (model inference); never present an estimate as measured; if a required metric is unavailable, mark it N/A — do not invent it.
Reference: See Ranking Analysis Templates for the complete output templates for all seven steps.
Example
A ranking report names the biggest winners, biggest drops, and next actions, with each position carrying its source tag and labeled delta against baseline.
Rank Change Quick Reference
Response Protocol
| Change |
Timeframe |
Action |
| Drop 1-3 positions |
Wait 1-2 weeks |
Monitor — may be normal fluctuation |
| Drop 3-5 positions |
Investigate within 1 week |
Check technical issues and competitor changes |
| Drop 5-10 positions |
Investigate immediately |
Run a full diagnostic: technical, content, links |
| Drop off page 1 |
Emergency response |
Comprehensive audit + recovery plan |
| Position gained |
Document and learn |
Identify what worked and replicate |
Reference: See Tracking Setup Guide for tracking setup, root-cause taxonomy, CTR benchmarks, SERP feature impact, and algorithm-update assessment.
Save Results
Ask "Save these results?" If yes, write to memory/monitoring/ — see Skill Contract §Save Results Template.
Reference Materials
- Tracking Setup Guide — Setup rules, feature tracking, and interpretation guidance
Next Best Skill
Initial setup (no baseline) → alert-manager. Subsequent runs (baseline exists) → Terminal. Visited-set rule applies per skill-contract.md.
1---2name: rank-tracker3description: Use when the user asks to "track rankings" or "查排名"; measures keyword and SERP-position deltas over time from provided exports or connected tools, including AI-response checks. Not for multi-metric stakeholder reports — use performance-reporter; not for setting alerts — use alert-manager. 排名追踪/SERP监控4license: Apache-2.05---67# Rank Tracker89Tracks keyword positions, SERP feature ownership, and AI visibility over time.1011## Quick Start1213```14Set up rank tracking for [domain] targeting these keywords: [keyword list]15```1617```18Analyze ranking changes for [domain] over the past [time period]19```2021## Skill Contract2223**Expected output**: a ranking report or delta summary plus the standard handoff summary for `memory/monitoring/`.2425- **Reads**: current rankings, prior baselines, target keyword list, market/device, and any user-provided or tool metrics.26- **Writes**: a user-facing monitoring deliverable and reusable summary.27- **Promotes**: significant changes, confirmed anomalies, follow-up actions, and pending decisions to `memory/open-loops.md`.28- **Done when**: every tracked keyword shows current position vs baseline with a labeled delta (or N/A); each position cites its source (tool export / user-provided / estimated); and biggest movers and likely causes are named.29- **Primary next skill**: [alert-manager](../alert-manager/SKILL.md) when recurring monitoring should become automated.3031### Handoff Summary3233> Emit the standard shape from [skill-contract.md §Handoff Summary Format](../../references/skill-contract.md).3435## Data Sources3637All integrations optional (see [CONNECTORS.md](../../CONNECTORS.md)). With tools, pull rankings from ~~SEO tool, impressions from ~~search console, traffic from ~~analytics, and AI citations from ~~AI monitor. Without tools, ask for positions, volumes, competitor data, and SERP feature status.3839**Zero-dependency measurement loop** (no paid tool needed): never narrate a ranking movement you did not measure — record each check as a snapshot and let the ledger compute the delta. `python3 "${CLAUDE_PLUGIN_ROOT}/scripts/connectors/ledger.py" record <domain> --source rank --data '{"<keyword>": <position>, ...}'`, then `ledger.py diff <domain> --source rank` for the change since the last check and `ledger.py trend <domain> --source rank --field "<keyword>"` for the series. Record real Search Console positions when connected, or the positions the user provides at Tier 1. A ranking delta is a **week-scale, confounded outcome** — attribute it against a control (a holdout keyword/competitor), not against zero; see [references/measurement-protocol.md](../../references/measurement-protocol.md). See [scripts/connectors/README.md](../../scripts/connectors/README.md).4041## Decision Gates4243**Stop and ask the user when:**44- No target keywords are provided and none can be inferred from `CLAUDE.md` or prior monitoring records — offer: (1) supply a keyword list, (2) track the domain's top known terms, (3) cancel.4546**Continue silently (never stop for):**47- No prior baseline exists — record the current run as the baseline, label all positions as the first snapshot, and proceed (do not invent a "previous" position).48- Missing optional tool data (SERP features, AI citations) — mark N/A and proceed.4950## Instructions5152When a user requests rank tracking or analysis:53541. **Set Up Keyword Tracking** — configure domain, market, device, language, update frequency, priorities, and competitor watchlist.552. **Record Current Rankings** — output a position table where every row cites its source (tool export / user-provided / estimated), with position ranges, ranking URLs, feature ownership, and movement vs baseline.563. **Analyze Ranking Changes** — highlight biggest wins, declines, stable terms, new rankings, lost rankings, likely causes, and recovery ideas; each delta labeled against its baseline.574. **Track SERP Features** — compare ownership of snippets, PAA, image/video packs, local packs, and related feature shifts.585. **Track GEO / AI Visibility** — monitor AI Overview presence, citation rate, citation position, and trend; mark each value Measured (from an ~~AI monitor) or N/A if unobserved.596. **Compare Against Competitors** — report share of voice, head-to-head comparisons, and threat levels.607. **Generate Ranking Report** — output overall trend, key wins, concerns, opportunities, SERP feature changes, GEO visibility, and recommendations, with each metric carrying its source tag.6162Label every metric **Measured** (tool/export), **User-provided**, or **Estimated** (model inference); never present an estimate as measured; if a required metric is unavailable, mark it N/A — do not invent it.6364> **Reference**: See [Ranking Analysis Templates](references/ranking-analysis-templates.md) for the complete output templates for all seven steps.6566## Example6768A ranking report names the biggest winners, biggest drops, and next actions, with each position carrying its source tag and labeled delta against baseline.6970## Rank Change Quick Reference7172### Response Protocol7374| Change | Timeframe | Action |75|--------|-----------|--------|76| Drop 1-3 positions | Wait 1-2 weeks | Monitor — may be normal fluctuation |77| Drop 3-5 positions | Investigate within 1 week | Check technical issues and competitor changes |78| Drop 5-10 positions | Investigate immediately | Run a full diagnostic: technical, content, links |79| Drop off page 1 | Emergency response | Comprehensive audit + recovery plan |80| Position gained | Document and learn | Identify what worked and replicate |8182> **Reference**: See [Tracking Setup Guide](references/tracking-setup-guide.md) for tracking setup, root-cause taxonomy, CTR benchmarks, SERP feature impact, and algorithm-update assessment.8384### Save Results8586Ask "Save these results?" If yes, write to `memory/monitoring/` — see [Skill Contract](../../references/skill-contract.md) §Save Results Template.8788## Reference Materials8990- [Tracking Setup Guide](references/tracking-setup-guide.md) — Setup rules, feature tracking, and interpretation guidance9192## Next Best Skill9394Initial setup (no baseline) → [alert-manager](../alert-manager/SKILL.md). Subsequent runs (baseline exists) → Terminal. Visited-set rule applies per [skill-contract.md](../../references/skill-contract.md).
tuyv