WEPR GEO Suite
Turn a GEO objective into an evidence-led delivery chain. Route focused requests directly; orchestrate only the skills required for a multi-stage project.
Start with scope
- Record the brand, business goal, audience, market, platforms, period, budget, conversion action, available evidence, and required files.
- Separate public evidence, user-provided evidence, platform samples, assumptions, and unavailable data.
- Never represent synthetic prompts, inferred competitors, or generated examples as real platform data.
- Define success with observable indicators. Do not promise ranking, citations, recommendations, leads, or revenue.
Read references/routing-and-delivery.md for the skill map, stage gates, and standard project combinations.
Route the work
- Baseline and opportunity map:
$wepr-geo-panorama-audit.
- Natural-language query expansion and prompt library:
$wepr-geo-intent-miner.
- Brand facts and reusable evidence:
$wepr-geo-knowledge-base-builder; use $wepr-geo-brand-graph when entity relationships are central.
- Existing-page diagnosis:
$wepr-geo-page-audit; new page structure: $wepr-geo-page-blueprint.
- Content:
$wepr-geo-title-optimizer, $wepr-geo-explainer-builder, $wepr-geo-comparison-builder, $wepr-geo-ranking-article-builder, or $wepr-geo-content-refiner according to format.
- Platform evidence collection:
$wepr-deepseek-crawler, $wepr-doubao-crawler, or $wepr-chatgpt-crawler. Respect login, rate, CAPTCHA, privacy, and platform boundaries.
- Monitoring and attribution:
$wepr-geo-effect-monitor for answer/citation signals; $wepr-geo-tracking for downstream conversion attribution.
- 30/60/90-day implementation:
$wepr-geo-execution-roadmap.
- GEOFlow product development, system operations, frontend themes, channel sites, legacy migration, and client implementation:
$wepr-geoflow.
- Broader WEPR channel strategy, GEO/SEO integration, and commercial growth planning:
$plan-organic-growth; live GEOrank operations: $operate-georank-workbench.
- Proposal or report presentation:
$wepr-presentation-workbench; it selects editable PPTX or single-file interactive HTML according to the delivery requirement.
Orchestrate multi-stage projects
Default sequence:
- Establish evidence and measurement boundaries.
- Build a baseline and intent universe.
- Create the fact base and entity model.
- Prioritize pages and content by business value, evidence readiness, and platform fit.
- Produce or revise assets with source-bound claims.
- Sample target platforms under a documented protocol.
- Track answer quality, citations, conversions, and confounders.
- Convert findings into a 30/60/90-day roadmap and quarterly review.
Skip stages only when the user already provides an equivalent, current artifact. Preserve handoff fields between skills: entity IDs, source IDs, prompt IDs, page IDs, platform, sample mode, date, confidence, owner, status, and acceptance criteria.
Delivery contract
For a client-facing project, return:
- executive summary and decision boundaries;
- source ledger and data-access statement;
- intent and platform scope;
- prioritized workstreams, owners, timeline, and dependencies;
- deliverables and acceptance tests;
- measurement baseline, sampling protocol, and attribution limits;
- risks, approvals, and unresolved inputs;
- file links for every generated artifact.
Use one canonical content structure for Markdown, HTML, Word, PDF, slides, and spreadsheets. Verify formulas, links, citations, file validity, layout overflow, and consistency before delivery.
Quality gates
- Keep brand facts traceable to sources and dates.
- Label fact, inference, estimate, synthetic example, and real sample distinctly.
- Use natural user questions, not mechanical suffix lists.
- Compare competitors on the same dimensions and preserve legitimate strengths.
- Treat platform metrics as repeated-sample estimates, not ground truth.
- Keep platform collection low-frequency, visible, authorized, and auditable.
- Require explicit approval before live publication, destructive operations, bulk synchronization, secret access, or production activation.
- Preserve applicable legal notices when redistributing packaged skills.
1---2name: wepr-geo-suite3description: 作为 WEPR GEO 项目的统一入口，负责组合调研、AI 搜索意图挖掘、全景诊断、知识资产、页面与内容生产、平台采样、效果监测、归因、实施路线图、GEOFlow 操作和客户交付。适用于跨多个工作模块、平台、季度或交付物的项目，也适用于暂时无法判断应调用哪个专项技能的情况。4---56# WEPR GEO Suite78Turn a GEO objective into an evidence-led delivery chain. Route focused requests directly; orchestrate only the skills required for a multi-stage project.910## Start with scope11121. Record the brand, business goal, audience, market, platforms, period, budget, conversion action, available evidence, and required files.132. Separate public evidence, user-provided evidence, platform samples, assumptions, and unavailable data.143. Never represent synthetic prompts, inferred competitors, or generated examples as real platform data.154. Define success with observable indicators. Do not promise ranking, citations, recommendations, leads, or revenue.1617Read [references/routing-and-delivery.md](references/routing-and-delivery.md) for the skill map, stage gates, and standard project combinations.1819## Route the work2021- Baseline and opportunity map: `$wepr-geo-panorama-audit`.22- Natural-language query expansion and prompt library: `$wepr-geo-intent-miner`.23- Brand facts and reusable evidence: `$wepr-geo-knowledge-base-builder`; use `$wepr-geo-brand-graph` when entity relationships are central.24- Existing-page diagnosis: `$wepr-geo-page-audit`; new page structure: `$wepr-geo-page-blueprint`.25- Content: `$wepr-geo-title-optimizer`, `$wepr-geo-explainer-builder`, `$wepr-geo-comparison-builder`, `$wepr-geo-ranking-article-builder`, or `$wepr-geo-content-refiner` according to format.26- Platform evidence collection: `$wepr-deepseek-crawler`, `$wepr-doubao-crawler`, or `$wepr-chatgpt-crawler`. Respect login, rate, CAPTCHA, privacy, and platform boundaries.27- Monitoring and attribution: `$wepr-geo-effect-monitor` for answer/citation signals; `$wepr-geo-tracking` for downstream conversion attribution.28- 30/60/90-day implementation: `$wepr-geo-execution-roadmap`.29- GEOFlow product development, system operations, frontend themes, channel sites, legacy migration, and client implementation: `$wepr-geoflow`.30- Broader WEPR channel strategy, GEO/SEO integration, and commercial growth planning: `$plan-organic-growth`; live GEOrank operations: `$operate-georank-workbench`.31- Proposal or report presentation: `$wepr-presentation-workbench`; it selects editable PPTX or single-file interactive HTML according to the delivery requirement.3233## Orchestrate multi-stage projects3435Default sequence:36371. Establish evidence and measurement boundaries.382. Build a baseline and intent universe.393. Create the fact base and entity model.404. Prioritize pages and content by business value, evidence readiness, and platform fit.415. Produce or revise assets with source-bound claims.426. Sample target platforms under a documented protocol.437. Track answer quality, citations, conversions, and confounders.448. Convert findings into a 30/60/90-day roadmap and quarterly review.4546Skip stages only when the user already provides an equivalent, current artifact. Preserve handoff fields between skills: entity IDs, source IDs, prompt IDs, page IDs, platform, sample mode, date, confidence, owner, status, and acceptance criteria.4748## Delivery contract4950For a client-facing project, return:5152- executive summary and decision boundaries;53- source ledger and data-access statement;54- intent and platform scope;55- prioritized workstreams, owners, timeline, and dependencies;56- deliverables and acceptance tests;57- measurement baseline, sampling protocol, and attribution limits;58- risks, approvals, and unresolved inputs;59- file links for every generated artifact.6061Use one canonical content structure for Markdown, HTML, Word, PDF, slides, and spreadsheets. Verify formulas, links, citations, file validity, layout overflow, and consistency before delivery.6263## Quality gates6465- Keep brand facts traceable to sources and dates.66- Label fact, inference, estimate, synthetic example, and real sample distinctly.67- Use natural user questions, not mechanical suffix lists.68- Compare competitors on the same dimensions and preserve legitimate strengths.69- Treat platform metrics as repeated-sample estimates, not ground truth.70- Keep platform collection low-frequency, visible, authorized, and auditable.71- Require explicit approval before live publication, destructive operations, bulk synchronization, secret access, or production activation.72- Preserve applicable legal notices when redistributing packaged skills.
ShenyuanNext