Release content operations 1.5.0
Browse files- README.md +27 -2
- SKILL.md +55 -6
- agents/openai.yaml +2 -2
- references/audience-first.md +36 -0
- references/channels.md +2 -0
- references/competitor-content.md +27 -0
- references/delivery.md +4 -0
- references/diagnosis.md +27 -0
- references/fields.md +15 -0
- references/intake.md +16 -0
- references/provenance.md +2 -0
- references/search-positioning.md +60 -0
- references/templates.md +72 -0
- references/website-to-social.md +33 -0
- references/worked-cases.md +54 -0
README.md
CHANGED
|
@@ -14,7 +14,7 @@ tags:
|
|
| 14 |
|
| 15 |
# AI Content Operations / AI 内容运营
|
| 16 |
|
| 17 |
-
An installable agent skill for
|
| 18 |
|
| 19 |
## Install
|
| 20 |
|
|
@@ -32,6 +32,31 @@ Outputs include a channel/capacity card, source register, weekly calendar, paren
|
|
| 32 |
|
| 33 |
This skill coordinates an operation. A single article adaptation belongs in a focused writing workflow. Publishing and direct messages require explicit authorization; unavailable metrics remain unknown.
|
| 34 |
|
| 35 |
-
Version: 1.
|
| 36 |
|
| 37 |
Source: https://github.com/Gingiris-1031/gingiris-skills/tree/main/skills/gr-ai-content-ops
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
|
| 15 |
# AI Content Operations / AI 内容运营
|
| 16 |
|
| 17 |
+
An installable agent skill for audience-first content, competitor research, SEO/GEO positioning, substantive drafts and measurable content operations. 面向受众价值,连接竞品研究、关键词定位、完整内容与运营复盘。
|
| 18 |
|
| 19 |
## Install
|
| 20 |
|
|
|
|
| 32 |
|
| 33 |
This skill coordinates an operation. A single article adaptation belongs in a focused writing workflow. Publishing and direct messages require explicit authorization; unavailable metrics remain unknown.
|
| 34 |
|
| 35 |
+
Version: 1.5.0. License: MIT.
|
| 36 |
|
| 37 |
Source: https://github.com/Gingiris-1031/gingiris-skills/tree/main/skills/gr-ai-content-ops
|
| 38 |
+
|
| 39 |
+
## Version 1.1.0
|
| 40 |
+
|
| 41 |
+
Adds stage-matched competitor content research, dated launch evidence, transferability checks and evidence-to-experiment planning. 新增竞品早期内容研究、历史证据与可迁移性判断,调研结论直接连接内容实验。
|
| 42 |
+
|
| 43 |
+
The planning workflow now asks a short adaptive intake for product and competitor URLs, audience and desired action, skipping known answers and supporting products without public websites. 规划前主动补问产品/竞品网址、受众与目标;已有信息不重复询问,未上线产品可提供简述。
|
| 44 |
+
|
| 45 |
+
## Version 1.2.0
|
| 46 |
+
|
| 47 |
+
Adds website-to-social planning, existing-content claim audits, informative visual briefs, independent asset readiness and activation measurement. Includes dated public-source Ring2 and Analook planning examples, not performance case studies. 新增官网到社媒的内容流程,以 Ring2 和 Analook 展示不同的素材、证据和转化路径;案例不是增长成效背书。
|
| 48 |
+
|
| 49 |
+
### 1.2.1
|
| 50 |
+
Complete LinkedIn case/analysis drafts with substantive reasoning; cooperative claim-level verification of user-confirmed material. 完整长稿与逐项辅助验证,避免把核验缺口误写成内容不实。
|
| 51 |
+
|
| 52 |
+
## Version 1.3.0
|
| 53 |
+
|
| 54 |
+
Adds a pre-draft SEO/GEO brief: intent clusters, localized SERP evidence, page ownership, differentiated content positioning, evidence-backed outlines and separate citation/conversion measurement. 博客动笔前明确关键词、搜索意图、内容定位、URL分工与差异化证据;候选词不冒充已验证机会。
|
| 55 |
+
|
| 56 |
+
## Version 1.4.0
|
| 57 |
+
|
| 58 |
+
Audience value precedes keyword and production planning: recognisable situations, competing angles, substantive examples and reader-centred editing. 先判断潜在受众为什么愿意看,再做关键词、格式和分发;编辑评分不冒充受众验证。
|
| 59 |
+
|
| 60 |
+
## Version 1.5.0
|
| 61 |
+
|
| 62 |
+
Adds retrieval-first diagnosis, competing explanations, experiment feasibility, persistent state and bilingual field definitions; preserves audience-first writing and SEO/GEO positioning. 新增取数优先、原因判别、实验可判定性、跨会话台账与缺失恢复,按任务选择模板。
|
SKILL.md
CHANGED
|
@@ -1,13 +1,14 @@
|
|
| 1 |
---
|
| 2 |
name: gr-ai-content-ops
|
| 3 |
description: >-
|
| 4 |
-
Build or audit an evidence-led AI content operation: source libraries, channel roles,
|
| 5 |
weekly calendars, editorial gates, repurposing, distribution receipts and business
|
| 6 |
measurement. Use when planning 内容运营、内容矩阵、周排期、多平台分发、素材库 and content workflow
|
| 7 |
-
recovery
|
|
|
|
| 8 |
license: MIT
|
| 9 |
metadata:
|
| 10 |
-
version: "1.
|
| 11 |
author: Gingiris
|
| 12 |
---
|
| 13 |
|
|
@@ -15,6 +16,12 @@ metadata:
|
|
| 15 |
|
| 16 |
Turn real material into a manageable content operation whose delivery and outcomes can be checked. Work in the user's language. Deliver the requested scope; a single-platform team does not need a multi-channel matrix.
|
| 17 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
## Start with the constraint
|
| 19 |
|
| 20 |
Use supplied context first. Resolve only missing decisions that change the work:
|
|
@@ -24,7 +31,31 @@ Use supplied context first. Resolve only missing decisions that change the work:
|
|
| 24 |
- Available evidence, publishing rights, existing channels, recent results and weekly capacity.
|
| 25 |
- Requested endpoint: diagnosis, weekly plan, draft pack, authorized distribution, or review.
|
| 26 |
|
| 27 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 28 |
|
| 29 |
## 1. Choose a content engine and channel roles
|
| 30 |
|
|
@@ -34,6 +65,18 @@ Select a primary channel using observed acquisition and audience behavior. Possi
|
|
| 34 |
|
| 35 |
Assign each chosen channel one job, a suitable format and a next step. Distinguish founder and company accounts so they have a reason to coexist. Avoid adding channels before the present workflow fits capacity. See [channel decisions](references/channels.md).
|
| 36 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
## 2. Build a usable source queue
|
| 38 |
|
| 39 |
Preserve raw material separately from publication-ready extracts. For each extract record source, date, exact locator, speaker, claim, verification state and allowed audience. A private transcript is not public evidence; anonymization alone does not settle permission or reidentification risk.
|
|
@@ -50,6 +93,12 @@ Choose topics with a specific audience, a supportable point and a useful action.
|
|
| 50 |
|
| 51 |
Use absolute dates and time zones, an owner, reviewer, estimate, evidence link and delivery state. Keep parent/variant identifiers and separate domestic and international language/audience needs where relevant. If over capacity, reduce scope or cadence before adding people. Use [templates](references/templates.md).
|
| 52 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
## 4. Produce through three gates
|
| 54 |
|
| 55 |
1. **Evidence:** check material factual claims and quotations; resolve unsupported claims, label uncertainty or remove them. One authoritative source may be enough; a second source is useful where claims are disputed or weak.
|
|
@@ -78,13 +127,13 @@ Separate content performance, account development and business outcomes. Specify
|
|
| 78 |
|
| 79 |
Compare typical performance using medians alongside sample sizes, means and outliers when useful. Isolate paid amplification and activity-driven cohorts where possible. Trace attributable business actions with available campaign links, codes or user reports, while disclosing attribution gaps.
|
| 80 |
|
| 81 |
-
End with the supported decision: continue, change, stop
|
| 82 |
|
| 83 |
## Deliverables and completion
|
| 84 |
|
| 85 |
Select only the artifacts needed: channel diagnosis, source register, topic queue, weekly calendar, reviewed parent and variants, distribution ledger, or review memo. Use the copyable fields in [templates](references/templates.md).
|
| 86 |
|
| 87 |
-
A plan is complete when work fits capacity and every item has an owner, evidence state and next step. A draft is complete when reviews and media checks are recorded. Distribution is complete only for targets with verified receipts; failed, unknown or unapproved targets remain explicit. A review is complete when its evidence window, gaps and next decision are stated.
|
| 88 |
|
| 89 |
## Boundaries with related workflows
|
| 90 |
|
|
|
|
| 1 |
---
|
| 2 |
name: gr-ai-content-ops
|
| 3 |
description: >-
|
| 4 |
+
Build or audit an evidence-led AI content operation: competitor research, SEO/GEO keyword and content positioning, source libraries, channel roles,
|
| 5 |
weekly calendars, editorial gates, repurposing, distribution receipts and business
|
| 6 |
measurement. Use when planning 内容运营、内容矩阵、周排期、多平台分发、素材库 and content workflow
|
| 7 |
+
recovery, 选题、内容日历、内容SOP、一鱼多吃、内容复盘 and publishing receipts.
|
| 8 |
+
For a single supplied article rewrite, keep the scope to that deliverable.
|
| 9 |
license: MIT
|
| 10 |
metadata:
|
| 11 |
+
version: "1.5.0"
|
| 12 |
author: Gingiris
|
| 13 |
---
|
| 14 |
|
|
|
|
| 16 |
|
| 17 |
Turn real material into a manageable content operation whose delivery and outcomes can be checked. Work in the user's language. Deliver the requested scope; a single-platform team does not need a multi-channel matrix.
|
| 18 |
|
| 19 |
+
## Start with the reader's reason to care
|
| 20 |
+
|
| 21 |
+
The primary editorial goal is useful, engaging content for the intended reader. Workflow completion, word count, keywords and publishing volume are supporting constraints, not evidence of good content. Before choosing a format or drafting, identify a concrete reader situation, what is at stake, and the new understanding or useful outcome the piece delivers. Use actual customer questions, sales/support conversations, relevant audience discussions and observed content responses where available. Label inferred audience needs; don't pretend an imagined persona has validated them.
|
| 22 |
+
|
| 23 |
+
Use the [audience-first editorial gate](references/audience-first.md). Draft competing angles, choose the one with the clearest reader value and specific support, then write. SEO should help the right reader discover that answer; it should not turn the piece into generic advice. A demo should reveal behavior the reader cares about, not simply display a feature.
|
| 24 |
+
|
| 25 |
## Start with the constraint
|
| 26 |
|
| 27 |
Use supplied context first. Resolve only missing decisions that change the work:
|
|
|
|
| 31 |
- Available evidence, publishing rights, existing channels, recent results and weekly capacity.
|
| 32 |
- Requested endpoint: diagnosis, weekly plan, draft pack, authorized distribution, or review.
|
| 33 |
|
| 34 |
+
### Ask a short, adaptive intake
|
| 35 |
+
|
| 36 |
+
For strategy or planning, proactively ask for missing context that would materially improve the result. Start with at most three short questions, in the user's language; skip anything already supplied:
|
| 37 |
+
|
| 38 |
+
1. What is the product website (or a short description if it is not public yet)?
|
| 39 |
+
2. Which 1–2 competitors or reference brands should we study? Share URLs if available, and whether the interest is their early launch or current content. “Not sure” is a valid answer.
|
| 40 |
+
3. Who should the content reach, and what is the priority action this month: awareness, trial, demo requests or qualified leads?
|
| 41 |
+
|
| 42 |
+
Use the [intake template](references/templates.md#0-adaptive-intake) as phrasing guidance, not a mandatory questionnaire. Ask a later targeted question about geography/language, existing accounts, usable proof or weekly capacity only when it changes a decision. Do not make the user fill out every field before starting.
|
| 43 |
+
|
| 44 |
+
While awaiting answers, inspect supplied public sources and continue independent work. If competitors are unknown, propose candidates based on buyer and use case, label them provisional and explain the fit. If a URL is inaccessible, ask for a product description or accessible source; never claim to have audited it. If the user wants an immediate plan, state assumptions and proceed, retaining unresolved choices in the handoff. Do not finalize dependent choices while a necessary clarification is pending. User-provided facts, website claims, verified observations and planning assumptions remain distinct.
|
| 45 |
+
|
| 46 |
+
Before treating analytics or history as unavailable, check accessible exports, connected tools and existing project records. Distinguish not retrieved, access blocked, not instrumented and genuinely absent. For a diagnosis, retrieve the cheapest discriminating fields before proposing a long experiment; use [diagnosis](references/diagnosis.md). For a new account with no history, produce a baseline plan. Do not invent performance or claim that an account was audited. A user's request to plan or draft does not authorize sending messages or publishing posts.
|
| 47 |
+
|
| 48 |
+
## Before choosing channels: learn from comparable competitors
|
| 49 |
+
|
| 50 |
+
For a new strategy, first-month plan, or requested competitor benchmark, research 1–2 comparable products before fixing topics and cadence. Use a dated early-stage window when studying cold start; current content is not evidence of launch strategy. A supplied single-article rewrite does not need this step.
|
| 51 |
+
|
| 52 |
+
Connect dated original assets to audience, message, proof, CTA and accessible response evidence. Distinguish observed content from self-reported results and causal explanations. Assess differences in buyer, stage, product readiness and distribution resources before transferring a tactic. Turn each useful finding into a specific plan change and a measurable experiment; do not deliver only a competitor inventory. If historical evidence is inaccessible, disclose the gap and proceed with an explicit hypothesis.
|
| 53 |
+
|
| 54 |
+
Use [competitor content research](references/competitor-content.md) and its evidence card. Keep this workflow standalone; an installed `gr-competitor` may support deeper research when needed.
|
| 55 |
+
|
| 56 |
+
## Initialize only the operating state the task needs
|
| 57 |
+
|
| 58 |
+
For recurring planning, delivery or diagnosis, reuse the existing project workspace or establish a named local output directory. Record the path, current revision, evidence register and delivery ledger using [fields and persistence](references/fields.md). A one-off rewrite needs only its source and draft, not a complete operations database. Missing material preparation becomes a scheduled production task; missing names become unassigned roles with a handoff deadline. See [intake and recovery](references/intake.md).
|
| 59 |
|
| 60 |
## 1. Choose a content engine and channel roles
|
| 61 |
|
|
|
|
| 65 |
|
| 66 |
Assign each chosen channel one job, a suitable format and a next step. Distinguish founder and company accounts so they have a reason to coexist. Avoid adding channels before the present workflow fits capacity. See [channel decisions](references/channels.md).
|
| 67 |
|
| 68 |
+
### Choose and audit the parent asset before setting cadence
|
| 69 |
+
|
| 70 |
+
For a website-backed content plan, inventory existing guides, comparison pages, reports and demos before proposing new articles. Choose update, create, consolidate or defer by buyer intent, evidence readiness and conversion path. A blog-first plan is appropriate when a durable answer supports discovery and evaluation; demo-first or social-first may be better when comprehension or rapid feedback is the bottleneck. Do not make four blogs/eight posts a universal quota.
|
| 71 |
+
|
| 72 |
+
Audit existing owned content before repurposing it: publication does not validate statistics, first-person stories, comparative claims or onboarding promises. Resolve contradictions, add evidence or remove unsupported claims. Map each selected parent to a standalone social argument, an informative visual and a verified next step; keep internal review notes out of publishable copy. See [website-to-social workflow](references/website-to-social.md) and [two worked cases](references/worked-cases.md).
|
| 73 |
+
|
| 74 |
+
### Establish the search and content position before drafting blogs
|
| 75 |
+
|
| 76 |
+
For search-oriented blog planning/writing, produce a compact [SEO/GEO positioning brief](references/search-positioning.md) before drafting. Identify the buyer question, target market/language, primary intent cluster, supporting queries/entities, search-result page types, existing URL ownership, unique evidence/angle and next business action. Content position means the role and differentiated promise of the article, not merely its target ranking.
|
| 77 |
+
|
| 78 |
+
Verify keywords with available first-party query data, current localized search results and keyword tools. Record sources/dates and distinguish observed demand from candidate phrases; unknown volume, difficulty or AI visibility is not zero. Do not force all intents into blogs or create new URLs for synonymous queries. Connect each section to a reader question and useful evidence, then map the article to native social angles. Missing keyword-tool access permits a provisional brief and draft, not invented metrics or an optimization-complete claim.
|
| 79 |
+
|
| 80 |
## 2. Build a usable source queue
|
| 81 |
|
| 82 |
Preserve raw material separately from publication-ready extracts. For each extract record source, date, exact locator, speaker, claim, verification state and allowed audience. A private transcript is not public evidence; anonymization alone does not settle permission or reidentification risk.
|
|
|
|
| 93 |
|
| 94 |
Use absolute dates and time zones, an owner, reviewer, estimate, evidence link and delivery state. Keep parent/variant identifiers and separate domestic and international language/audience needs where relevant. If over capacity, reduce scope or cadence before adding people. Use [templates](references/templates.md).
|
| 95 |
|
| 96 |
+
### Write to substantive depth
|
| 97 |
+
|
| 98 |
+
For LinkedIn analysis and case posts, deliver a complete native argument, not a short teaser or a stretched X post. Include a concrete situation, relevant evidence/example, interpretation, practical action and a relevant next step. As a drafting heuristic, 180–300 English words often provides room; adapt to substance and current editor limits rather than enforcing a quota. Add depth through decisions and examples, not repeated disclaimers. For blog drafts, deliver the full requested article rather than an outline labeled as a draft.
|
| 99 |
+
|
| 100 |
+
Treat user-confirmed owned material as usable first-party source material. Assist with claim-level checks of freshness, units, sources and product behavior; 'not independently verified' does not mean false. Preserve supported detail and authentic narrative instead of replacing the whole case with generic advice. Separate user confirmation, published-source consistency and independent verification in internal notes. Narrow unresolved checks to affected claims; do not block unrelated copy. Keep only reader-relevant qualifiers in the post and move production checks to a separate review sheet.
|
| 101 |
+
|
| 102 |
## 4. Produce through three gates
|
| 103 |
|
| 104 |
1. **Evidence:** check material factual claims and quotations; resolve unsupported claims, label uncertainty or remove them. One authoritative source may be enough; a second source is useful where claims are disputed or weak.
|
|
|
|
| 127 |
|
| 128 |
Compare typical performance using medians alongside sample sizes, means and outliers when useful. Isolate paid amplification and activity-driven cohorts where possible. Trace attributable business actions with available campaign links, codes or user reports, while disclosing attribution gaps.
|
| 129 |
|
| 130 |
+
For a diagnosis, compare plausible causes, name the evidence that would weaken each, and prefer a cheap discriminating check. Use [diagnosis](references/diagnosis.md) before proposing an experiment. End with the supported decision: continue, change, stop, gather missing evidence or inconclusive; assign an owner and next review date. A future check is a scheduled task only if a real scheduler was configured with authorization. Otherwise deliver a dated checklist and say it has not been scheduled.
|
| 131 |
|
| 132 |
## Deliverables and completion
|
| 133 |
|
| 134 |
Select only the artifacts needed: channel diagnosis, source register, topic queue, weekly calendar, reviewed parent and variants, distribution ledger, or review memo. Use the copyable fields in [templates](references/templates.md).
|
| 135 |
|
| 136 |
+
A plan is complete when work fits capacity and every item has an owner (or an unassigned role with an assignment deadline), evidence state and next step. A draft is complete when reviews and media checks are recorded. Distribution is complete only for targets with verified receipts; failed, unknown or unapproved targets remain explicit. A review is complete when its evidence window, gaps and next decision are stated.
|
| 137 |
|
| 138 |
## Boundaries with related workflows
|
| 139 |
|
agents/openai.yaml
CHANGED
|
@@ -1,4 +1,4 @@
|
|
| 1 |
interface:
|
| 2 |
display_name: "AI Content Operations"
|
| 3 |
-
short_description: "Plan
|
| 4 |
-
default_prompt: "Use $gr-ai-content-ops to
|
|
|
|
| 1 |
interface:
|
| 2 |
display_name: "AI Content Operations"
|
| 3 |
+
short_description: "Plan audience-first content, SEO briefs and measurable reviews"
|
| 4 |
+
default_prompt: "Use $gr-ai-content-ops to identify what my audience wants to read, research relevant competitors, and build a feasible content plan with SEO positioning, substantive drafts and measurable next steps."
|
references/audience-first.md
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Audience-first editorial gate
|
| 2 |
+
|
| 3 |
+
## Find a reason to read before filling a calendar
|
| 4 |
+
|
| 5 |
+
For each topic, answer in plain language:
|
| 6 |
+
- Who encounters this situation, and at what moment?
|
| 7 |
+
- What are they trying to gain, avoid, understand or decide?
|
| 8 |
+
- What concrete observation, example or explanation will reward their attention?
|
| 9 |
+
- Why can this author/product contribute something credible here?
|
| 10 |
+
|
| 11 |
+
Find reader language in available customer questions, support/sales records, relevant public conversations or consented interviews. Preserve source and context internally. User-confirmed experience is valid first-party material; assist with specific checks. An inferred scenario can be written as an illustration, not a reported customer incident. If evidence is thin, write a useful small piece or ask for one missing example instead of padding a universal framework.
|
| 12 |
+
|
| 13 |
+
## Select an angle
|
| 14 |
+
|
| 15 |
+
Develop 2–3 different angles before writing: a costly misunderstanding, a concrete decision with tradeoffs, a revealing demonstration, an overlooked detail, or an earned lesson. These are options, not mandatory conflict/contrarian formulas. State the reader payoff for each and choose based on specificity, relevance and evidence. Do not invent urgency, losses, consensus or controversy.
|
| 16 |
+
|
| 17 |
+
A topic such as 'AI handoff best practices' is not yet an angle. 'The morning employee has a transcript but still cannot return the call' supplies a situation and a practical question, provided it is framed as an illustration unless sourced.
|
| 18 |
+
|
| 19 |
+
## Write for attention and reward it
|
| 20 |
+
|
| 21 |
+
Open with the actual situation, observation or consequential question. Advance the thought: show a concrete detail, explain why it changes a decision, explore the relevant tradeoff and deliver a useful next step. Choose the structure that suits the material; do not mechanically repeat scene-conflict-lesson for every post.
|
| 22 |
+
|
| 23 |
+
Replace generic terms with things readers recognise: a transcript, a missing address, a price question, a calendar entry, the colleague returning a call. Explain product value in the reader's work, without making unsupported capability claims. A checklist belongs where it helps the reader act, not as a substitute for reasoning. Remove repeated warnings, empty rhetorical questions and unnecessary brand mentions. First-person experience belongs only to the actual speaker.
|
| 24 |
+
|
| 25 |
+
Choose a visual that contains useful information or demonstrates the point. A diagram brief is not a finished asset. The final CTA should follow from the reader's problem; an educational post need not carry a sales link if it adds nothing.
|
| 26 |
+
|
| 27 |
+
## Review before delivery
|
| 28 |
+
|
| 29 |
+
Read the draft as the intended buyer, not as an SEO auditor:
|
| 30 |
+
1. Is the opening recognisably relevant, or could it introduce any SaaS product?
|
| 31 |
+
2. Does the reader learn something concrete beyond 'define goals, measure and iterate'?
|
| 32 |
+
3. Does each paragraph develop the point rather than repeat it?
|
| 33 |
+
4. Does the evidence support the conclusion and the author's perspective?
|
| 34 |
+
5. Would the content still be useful without its CTA?
|
| 35 |
+
|
| 36 |
+
If it fails, revise the angle/example before adjusting length or keywords. Editorial review is a proxy, not proof readers like it. Validate through real target-reader feedback and available saves, substantive replies, shares, qualified clicks or business actions. Record sample/window and distinguish reactions from buyer fit. Use findings to choose the next topic; do not promise virality from a rubric or fabricate reader-panel feedback.
|
references/channels.md
CHANGED
|
@@ -25,3 +25,5 @@ Company account: explain capability, changes and proof. Founder/employee account
|
|
| 25 |
A reviewed tutorial might become a short viewpoint, a card set and a demonstration script. Each variant has its own hook, length, evidence mapping, next action and delivery state. Seven formats are a possibility, not a quota. Reuse can span weeks if the claim is still current and the audience/context is distinct.
|
| 26 |
|
| 27 |
Keep campaign messaging consistent between social and the relevant landing page. Changing the website requires scope and authorization; do not update it merely because a content plan mentions alignment.
|
|
|
|
|
|
|
|
|
| 25 |
A reviewed tutorial might become a short viewpoint, a card set and a demonstration script. Each variant has its own hook, length, evidence mapping, next action and delivery state. Seven formats are a possibility, not a quota. Reuse can span weeks if the claim is still current and the audience/context is distinct.
|
| 26 |
|
| 27 |
Keep campaign messaging consistent between social and the relevant landing page. Changing the website requires scope and authorization; do not update it merely because a content plan mentions alignment.
|
| 28 |
+
|
| 29 |
+
For website-backed campaigns and LinkedIn adaptations, use [website-to-social](website-to-social.md); compare the [Ring2 and Analook planning cases](worked-cases.md).
|
references/competitor-content.md
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Competitor content research / 竞品内容研究
|
| 2 |
+
|
| 3 |
+
## Scope before search
|
| 4 |
+
|
| 5 |
+
Choose 1–2 comparators by buyer, job and business model; a user-named competitor is a starting point, not proof of identical positioning. Define the decision the research must change. For cold start, anchor a dated public launch and a bounded period (for example the next 30–90 days). Call it an observed launch window, not the first-ever activity unless established. Date of retrieval and date of publication are separate.
|
| 6 |
+
|
| 7 |
+
## Reconstruct assets and their path to action
|
| 8 |
+
|
| 9 |
+
Prefer contemporaneous founder/company posts, launch threads, dated release notes, videos, documentation/repository history and archived pages. Read comments where they reveal actual questions or adoption friction. Search coverage is a sample; missing search results do not prove a channel was unused. Modern navigation and updated pages must not be attributed to their older article date. Current repository HEAD is not a historical README.
|
| 10 |
+
|
| 11 |
+
Record for each useful asset: date and locator; target buyer; problem/hook; format and demonstration; next action and friction; founder/company/community roles; response evidence; uncertainty. Check whether a linked demo/tutorial is merely referenced or actually accessible/tested. Do not run paid calls or contact people just to audit content.
|
| 12 |
+
|
| 13 |
+
Distinguish observed publication, company self-report, audience comment and analyst inference. Current likes/followers are not launch-day metrics. Keep paid/organic mix and conversion unknown unless sourced. Attention does not establish revenue causation. Include negative feedback, not only successful assets. Research methods and original paraphrases; do not copy competitor wording, testimonials or media.
|
| 14 |
+
|
| 15 |
+
## Transfer methods, not advantages
|
| 16 |
+
|
| 17 |
+
For each tactic assess buyer fit, product readiness, founder reach, partner/investor amplification, budget and rights. Mark adopt, adapt, defer or reject, with a reason. A developer API launch and an operator-facing service may share a demo-first method but require different channels and explanation. Investor promotion is a resource dependency, not a universal playbook step.
|
| 18 |
+
|
| 19 |
+
Convert findings into a parent asset, channel role, CTA, owner, time estimate, readiness gate and review metric. When proof/demo is unavailable, plan its production and hold claims until verified. Use a clearly labeled storyboard or requirements discussion as a fallback; never present it as a tested product or customer result. Do not default to high-volume SEO because a mature competitor has a large blog today.
|
| 20 |
+
|
| 21 |
+
## Evidence-to-action card
|
| 22 |
+
|
| 23 |
+
| Source and publication date | Historical window / coverage | Audience + problem | Asset + CTA | Observed response / self-report / unknown | Transfer constraints | Adopt/adapt/defer/reject | Plan change + owner | Experiment + review date |
|
| 24 |
+
|---|---|---|---|---|---|---|---|---|
|
| 25 |
+
| | | | | | | | | |
|
| 26 |
+
|
| 27 |
+
Completion: explain which planned items changed because of evidence, which did not, and what remains unknown. A useful small sample with stated limits is better than a fabricated comprehensive history.
|
references/delivery.md
CHANGED
|
@@ -37,3 +37,7 @@ For contributors, prepare a message including their approved asset, credit, link
|
|
| 37 |
Compare like windows, formats and audiences; split organic/paid when available. Report sample size, median and meaningful outliers. A high max/median ratio can motivate investigation but does not prove luck, causality or a broken strategy. Historical title analyses and internal rates are account-specific observations, not universal benchmarks.
|
| 38 |
|
| 39 |
Review memo: observation → plausible explanation → uncertainty → next controlled change → owner → review date. Do not change several factors and claim the observed result isolates one cause.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
Compare like windows, formats and audiences; split organic/paid when available. Report sample size, median and meaningful outliers. A high max/median ratio can motivate investigation but does not prove luck, causality or a broken strategy. Historical title analyses and internal rates are account-specific observations, not universal benchmarks.
|
| 38 |
|
| 39 |
Review memo: observation → plausible explanation → uncertainty → next controlled change → owner → review date. Do not change several factors and claim the observed result isolates one cause.
|
| 40 |
+
|
| 41 |
+
## Persistent state and receipt ownership
|
| 42 |
+
|
| 43 |
+
Use [fields](fields.md) for saved ledger paths/keys and [intake](intake.md) for receipt timing and owners. Authorization scopes are separate: plan/draft, upload a draft, schedule/publish specified content, and contact others. Prior explicit authorization carries forward within scope; these are not mandatory repeated approval steps. Report only completed receipts, with unexecuted targets clearly labeled.
|
references/diagnosis.md
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Diagnose before prescribing / 先诊断再开方案
|
| 2 |
+
|
| 3 |
+
## Retrieve before marking unknown
|
| 4 |
+
|
| 5 |
+
Check available project files, authorized connectors and backend exports before asking the user to supply data. Record field, location, window, availability, owner and next action. If retrieval requires access, request that missing access without asking for secrets in chat; progress independent work. Distinguish not-retrieved, access-blocked, not-instrumented, absent and unavailable-attempted. Retrieval cost is an estimate, not a universal two-hour promise.
|
| 6 |
+
|
| 7 |
+
Collect only fields that affect the decision: exposure/delivery denominator, source mix, meaningful reading or interactions, destination visits and business events. Platform definitions differ. A low read count could reflect fewer deliveries, a weak opening, a mismatch in audience or a tracking issue. A memory estimate can motivate checks but cannot substitute for an export.
|
| 8 |
+
|
| 9 |
+
## Compare causes and discriminate cheaply
|
| 10 |
+
|
| 11 |
+
| Hypothesis / 原因假设 | Expected signal / 预期信号 | Discriminating field / 判别字段 | What weakens it / 反证 | Access and cost / 获取成本 | Next check / 下一步 |
|
| 12 |
+
|---|---|---|---|---|---|
|
| 13 |
+
| Distribution fell | Lower deliveries or qualified reach | Delivery count/source mix | Reach stable with comparable cohort | Check existing export | Compare matched windows |
|
| 14 |
+
| Opening does not connect | Exposure stable, opening/click response lower | Platform-defined response rate | Response stable, drop occurs later | Check post detail | Inspect hook and audience fit |
|
| 15 |
+
| Promise and destination mismatch | Relevant visits, little next-step activity | Landing events, feedback | Visits never arrive or event is broken | Inspect analytics and flow | Check instrumentation and reader task |
|
| 16 |
+
|
| 17 |
+
These are possible concurrent causes, not forced mutually exclusive categories. Use as many hypotheses as evidence warrants. Check measurement definitions first; if exposure falls, inspect distribution before blaming the article. If exposure holds but engagement falls, examine audience/packaging with comparable samples. If visits arrive but conversion falls, inspect destination, offer and instrumentation. These signals narrow inquiry; none proves causation.
|
| 18 |
+
|
| 19 |
+
## Choose a test that can answer the question
|
| 20 |
+
|
| 21 |
+
Define the decision, unit of analysis, baseline, primary outcome, meaningful effect, expected sample, window and confounders. Posts/users are not interchangeable independent units. For a formal A/B test, justify sample size/power or detectable effect using an appropriate method and stated assumptions. If no defensible baseline/sample exists, use exploratory feedback or baseline collection rather than fabricated precision. A small sample or a 1.4x difference is inconclusive without uncertainty/context; it is not automatically proof or noise.
|
| 22 |
+
|
| 23 |
+
Prefer one interpretable change where feasible. When several changes are necessary, record them and limit attribution. Instrumentation repair can change comparability even if it is not the content treatment. Inconclusive is valid when data cannot discriminate: record the decision affected, missing evidence and cheapest next step. Do not require a fictional power calculation to admit uncertainty.
|
| 24 |
+
|
| 25 |
+
## Memo / 诊断交付
|
| 26 |
+
|
| 27 |
+
Observation + source/window → competing explanations → discriminating checks and results → supported decision or inconclusive → owner/deadline. For reader appeal, add the actual audience question and what the draft rewards the reader with. Operational metrics do not replace editorial judgment or target-reader feedback.
|
references/fields.md
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Fields and persistence / 字段与持久化
|
| 2 |
+
|
| 3 |
+
Keep three independent fields: evidence_status (verified / attributed / pending / unsupported), permission_status (allowed / unclear / restricted), delivery_status (source-needed / draft / review-needed / ready / authorized / submitted / scheduled / published-verified / measured; also blocked / failed / unknown / cancelled / draft-uploaded-verified). Verified evidence alone does not imply publication readiness. Properly attributed first-party material can be usable. Resolve material pending/unsupported claims by checking, qualifying or removing them; then reassess the draft. A published post can later need correction without erasing its delivery history.
|
| 4 |
+
|
| 5 |
+
For multi-session work, reuse the user's existing state store. Otherwise create a local task directory with README.md (scope, timezone, current version, next step), evidence.csv, calendar.csv and ledger.csv as needed. Record actual paths in the handoff. Store private raw material separately with limited access; never bundle it into a public skill. No new files are necessary for an ordinary single rewrite.
|
| 6 |
+
|
| 7 |
+
Ledger columns: operation_key, content_id, revision, platform, account, slot, timezone, authorization_scope, delivery_status, platform_id, verified_url, checked_at, owner, next_action. Key = content_id + platform + account + absolute slot. Revision is separate so editing copy does not bypass duplicate detection. A retry keeps the key. Rescheduling preserves the old row and requires applicable authorization. Check submitted/scheduled/published/unknown records before retry; the ledger is evidence of recorded work, not exhaustive platform history. Reconcile external changes when accessible.
|
| 8 |
+
|
| 9 |
+
Calendar fields: content_id, parent_id, date_time, timezone, platform, account, owner, reviewer, person_hours, evidence_link, asset_path, delivery_status, next_action. Capacity is summed person-hours, not elapsed time; show review and buffer separately. State an operational timezone, plus audience timezone when useful; handle daylight-saving conversion for dated slots.
|
| 10 |
+
|
| 11 |
+
Unassigned owner: unassigned(editor) plus assignment deadline. Untagged footage: to-be-tagged plus preparation task. Missing metrics: null/unknown plus reason, not zero. Missing permission: unclear plus scope/source to confirm, not an automatic conclusion based only on whether it was oral or written.
|
| 12 |
+
|
| 13 |
+
Platform-facts register (when a fact matters): platform | constraint | value | source/editor | checked_at | context | refresh_trigger. Recheck when stale, contradicted or editor behavior changes; do not require a fresh lookup of every stable fact every week.
|
| 14 |
+
|
| 15 |
+
中文对照:evidence_status=证据状态;permission_status=使用许可;delivery_status=交付状态;person_hours=人时;operation_key=操作键;verified_url=回读链接。Use the same field keys across exports; labels may be localized.
|
references/intake.md
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Missing inputs and recovery / 输入缺失与恢复
|
| 2 |
+
|
| 3 |
+
Use the short intake in SKILL.md; these are execution checks, not a mandatory opening questionnaire.
|
| 4 |
+
|
| 5 |
+
| Gap | Useful next action | Delivery implication |
|
| 6 |
+
|---|---|---|
|
| 7 |
+
| Analytics exist but were not supplied | Retrieve accessible exports or assign a specific field request | Diagnosis provisional; independent drafting can continue |
|
| 8 |
+
| No account history | Establish baseline collection | No invented best/worst post |
|
| 9 |
+
| Footage lacks timecodes, redaction or revision | Schedule tagging/redaction/version selection as a production task | Storyboard can be complete; video remains pending |
|
| 10 |
+
| Consent scope is unclear | Confirm the specific use, audience and material; record evidence | Hold affected story; use other authorized material |
|
| 11 |
+
| Oral permission reported | Assess explicit scope and reliable record; seek clarification if insufficient | Do not presume either authorization or prohibition solely from medium |
|
| 12 |
+
| Names missing | Assign a role placeholder and assignment deadline | Plan can be delivered with visible handoff task |
|
| 13 |
+
| Quote cannot be checked against audio | Inspect another reliable original; omit disputed words or use supported attributed paraphrase | Paraphrasing does not make an unsupported claim true |
|
| 14 |
+
| Platform access unavailable | Use reliable current documentation for planning, label editor untested | No claim of account audit/publication |
|
| 15 |
+
|
| 16 |
+
For receipts, define the operator, due time, location and user-facing summary. Agree on realistic timing: acceptance/readback during the authorized session when possible, metrics at the chosen window (for example 24h and 7d). If delayed or blocked, record attempted time and next step. A schedule only exists when configured, not when written in a plan. Draft-only requests receive a receipt template labeled not executed.
|
references/provenance.md
CHANGED
|
@@ -14,3 +14,5 @@ Generalization decisions:
|
|
| 14 |
- First-person style, title correlations and outlier ratios do not establish causal recommendations.
|
| 15 |
|
| 16 |
The skill is a reusable operating method, not a republication of the guide or a claim that its historical outcomes will repeat.
|
|
|
|
|
|
|
|
|
| 14 |
- First-person style, title correlations and outlier ratios do not establish causal recommendations.
|
| 15 |
|
| 16 |
The skill is a reusable operating method, not a republication of the guide or a claim that its historical outcomes will repeat.
|
| 17 |
+
|
| 18 |
+
Version 1.5.0 integrates an owner-supplied review of v1.0.0 with subsequent audience-first, competitor, long-form and search-positioning work. Retained retrieval-first diagnosis, persisted state, explicit gaps, template mapping and routing. Adapted suggested fixed thresholds, mutually-exclusive causes, rigid consent-medium rules and fixed receipt deadlines to context. The private review and test outputs are not bundled. Review observations about old releases are not current platform verification.
|
references/search-positioning.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# SEO/GEO keyword and content positioning brief
|
| 2 |
+
|
| 3 |
+
## 1. Define the reader and decision
|
| 4 |
+
|
| 5 |
+
State audience, market, language, product stage and intended business action. Ask for missing context only if it changes the plan. Write a positioning sentence: “For [reader] facing [decision], this page explains [answer] using [specific evidence], so they can [next step].” This is the content's promise, not the company's slogan or a ranking forecast.
|
| 6 |
+
|
| 7 |
+
## 2. Research queries and answer needs
|
| 8 |
+
|
| 9 |
+
Start with accessible Search Console queries/pages and customer questions; record the date range, geography, branded/non-branded split and missing coverage. Add competitor pages, search suggestions and current localized search results as discovery sources. Inspect a small representative set of actual results for the dominant page type, audience, intent, useful depth and unanswered questions. Record query, locale, check date and URLs; a snippet or search suggestion alone is not demand measurement.
|
| 10 |
+
|
| 11 |
+
Use available keyword tools for volume, difficulty and related queries, retaining provider, geography, period and retrieval date. Numbers from different providers are not directly comparable. Avoid universal volume/KD thresholds or guaranteed ranking claims. Low-volume, high-relevance questions can still merit a page. If tools are unavailable, explicitly mark candidate keywords and metric gaps, continue provisionally and do not call them validated opportunities.
|
| 12 |
+
|
| 13 |
+
Select one coherent primary intent cluster, a small useful set of secondary phrases/entities and natural-language buyer questions. Distinguish product competitors from search competitors. Do not stuff all phrase variants into headings or impose keyword density. Prioritize buyer relevance, actual opportunity evidence, differentiated proof, product fit and production cost; high volume alone does not set priority.
|
| 14 |
+
|
| 15 |
+
## 3. Assign page ownership and a differentiated position
|
| 16 |
+
|
| 17 |
+
Inventory existing pages before selecting create/update/consolidate/defer. Map equivalent intent to a preferred existing URL where sensible. Repeated words do not automatically mean cannibalization; inspect whether pages answer the same decision and, when available, whether query/page data suggests competition. Redirects or consolidation are proposed changes requiring normal implementation scope, not automatic actions.
|
| 18 |
+
|
| 19 |
+
Choose a guide for an explanatory task, a product/solution page for a capability or purchase task, and a comparison page for an actual evaluation. Do not force transactional searches into blog posts. Multi-brand work needs domain-level topic ownership, too.
|
| 20 |
+
|
| 21 |
+
Specify what this page contributes beyond current results: a verified workflow, original evidence, an explicit tradeoff, a worked example or a reusable decision tool. Generic completeness is not a differentiated position. Comparisons require dated, fairly scoped evidence for each product; 'best' and competitor names are not mandatory title ingredients.
|
| 22 |
+
|
| 23 |
+
## 4. Build the article and citation-ready evidence
|
| 24 |
+
|
| 25 |
+
Draft a clear title, H1, opening answer, section outline, proposed meta description and stable URL. Use the primary language naturally. Map each section to a reader question and the evidence needed. Include a useful visual and relevant internal links toward deeper explanation and a product next step. Preserve date, methodology and limitations beside original figures; distinguish first-party experience from independent findings.
|
| 26 |
+
|
| 27 |
+
For AI-answer discovery, describe entities consistently and provide concise, self-contained explanations, supported comparison tables or examples where useful. Natural-language questions help scope coverage; do not generate repetitive FAQ pages. Clear formatting and valid structured data do not guarantee inclusion or citation. Check current official guidance before asserting engine-specific requirements; there is no universal GEO file/schema requirement.
|
| 28 |
+
|
| 29 |
+
For an AI visibility check, record engine/model if available, exact prompt, date, locale/account context, answer evidence and cited URL. Separate brand mention, linked citation, factual accuracy and referral/conversion. An unchecked prompt list is a test plan, not a visibility result.
|
| 30 |
+
|
| 31 |
+
## 5. Hand off to social and measurement
|
| 32 |
+
|
| 33 |
+
Create native social arguments from the page's specific decision or finding, not from the keyword list. Specify visual, destination and tracking. Assess search impressions/clicks by query/page, relevant AI citations and meaningful business events separately. Record a baseline and review window; unavailable metrics remain unknown. Publishing, technical eligibility, indexing, ranking, citation and conversion each require their own evidence.
|
| 34 |
+
|
| 35 |
+
## Copyable brief
|
| 36 |
+
|
| 37 |
+
- Audience / market / language / business decision:
|
| 38 |
+
- Content positioning sentence:
|
| 39 |
+
- Primary intent cluster / candidate or verified:
|
| 40 |
+
- Supporting queries / entities / buyer questions:
|
| 41 |
+
- Keyword evidence: source, locale, date, volume/KD if available:
|
| 42 |
+
- SERP observations: query, date, locale, result URLs, dominant page type:
|
| 43 |
+
- Existing URL / competing internal pages / create-update-defer decision:
|
| 44 |
+
- Differentiated contribution / evidence or production dependency:
|
| 45 |
+
- Title / H1 / opening answer / proposed meta / URL:
|
| 46 |
+
- Outline: reader question → section → proof → useful visual:
|
| 47 |
+
- Internal links / product CTA / social adaptation:
|
| 48 |
+
- Baseline / outcome metric / review window / unresolved checks:
|
| 49 |
+
|
| 50 |
+
A brief can be complete as a provisional research artifact while demand verification remains pending. A final draft must fulfill the positioning promise and identify unverified claims in its separate production notes.
|
| 51 |
+
|
| 52 |
+
## Worked candidate positions (not keyword research results)
|
| 53 |
+
|
| 54 |
+
| Product | Candidate intent | Position and distinctive asset | Page ownership | Next action |
|
| 55 |
+
|---|---|---|---|---|
|
| 56 |
+
| Ring2 | virtual receptionist services; human vs AI receptionist | Help service operators choose responsibilities using a human/AI/hybrid workflow comparison | Inspect/update the existing reception article rather than create a near duplicate | Discuss one call workflow |
|
| 57 |
+
| Ring2 | AI call transfer; AI to human handoff | Explain what a completed handoff requires using an exception-path diagram | Inspect existing phone answering/transfer pages and split intent only if justified | Review system requirements |
|
| 58 |
+
| Analook | competitor SEO analysis; competitive intelligence for SEO | Help growth teams move from observed rival signals to a publishing decision using a worked evidence-to-action table | Update the existing competitor SEO guide | Read the relevant report then analyze a URL |
|
| 59 |
+
|
| 60 |
+
These phrases are editorial candidates. No search volume, difficulty, current ranking or AI citation has been established by this table.
|
references/templates.md
CHANGED
|
@@ -2,6 +2,28 @@
|
|
| 2 |
|
| 3 |
Populate only the requested scope. Blank and unknown are better than invented facts.
|
| 4 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
## 1. Channel and capacity card
|
| 6 |
|
| 7 |
- Audience / current understanding:
|
|
@@ -60,3 +82,53 @@ Populate only the requested scope. Blank and unknown are better than invented fa
|
|
| 60 |
| | | | | | | | | |
|
| 61 |
|
| 62 |
Name one reusable pattern and one bottleneck if the data supports them. Separate results from explanations. For no-history teams, report a baseline-collection plan instead of a fictional best/worst post.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
|
| 3 |
Populate only the requested scope. Blank and unknown are better than invented facts.
|
| 4 |
|
| 5 |
+
## 0. Adaptive intake
|
| 6 |
+
|
| 7 |
+
Ask only unanswered questions that affect this assignment. Example opening in Chinese:
|
| 8 |
+
|
| 9 |
+
> 为了让方案更贴近你的产品,先补充三点(已有的信息不用重复):
|
| 10 |
+
> 1. 产品官网是什么?还没上线的话,一句话介绍产品也可以。
|
| 11 |
+
> 2. 想参考哪1–2个竞品?可以给网址,并说说想看早期冷启动还是现在的内容;不确定的话我可以先找候选。
|
| 12 |
+
> 3. 这次主要面向谁,最希望内容带来什么行动:了解产品、试用、预约演示,还是有效咨询?
|
| 13 |
+
|
| 14 |
+
English equivalent: “What is your product URL or short description? Which competitors should we learn from, and at what stage? Who should we reach and what action should the content drive?”
|
| 15 |
+
|
| 16 |
+
Follow up only as needed: target market/language; existing account URLs and metrics; public demos or approved case material; hours available per week; delivery dates. A supplied article rewrite usually needs only its audience, destination and intended action.
|
| 17 |
+
|
| 18 |
+
Record the brief:
|
| 19 |
+
|
| 20 |
+
| Field | Answer / source | User-confirmed / observed / assumption / unknown | Decision affected / next step |
|
| 21 |
+
|---|---|---|---|
|
| 22 |
+
| Product URL or description | | | |
|
| 23 |
+
| Competitor URLs and historical window | | | |
|
| 24 |
+
| Audience and priority action | | | |
|
| 25 |
+
| Market, channels, proof, capacity (as needed) | | | |
|
| 26 |
+
|
| 27 |
## 1. Channel and capacity card
|
| 28 |
|
| 29 |
- Audience / current understanding:
|
|
|
|
| 82 |
| | | | | | | | | |
|
| 83 |
|
| 84 |
Name one reusable pattern and one bottleneck if the data supports them. Separate results from explanations. For no-history teams, report a baseline-collection plan instead of a fictional best/worst post.
|
| 85 |
+
|
| 86 |
+
## 9. Competitor evidence to content experiment
|
| 87 |
+
|
| 88 |
+
Use the [evidence-to-action card](competitor-content.md) for strategy and cold-start work. Add a one-line change log: previous plan → new plan → source → transfer constraint. Keep historical observations separate from proposed cadence and expected outcomes.
|
| 89 |
+
|
| 90 |
+
## 10. Parent-to-social production card
|
| 91 |
+
|
| 92 |
+
| Buyer question / intent | Existing URL | Update/create/defer + reason | Claims to verify | Parent asset | Native social argument | Visual brief + source + ready/pending | Verified destination + UTM | Owner / hours / review gate | Activation event + denominator |
|
| 93 |
+
|---|---|---|---|---|---|---|---|---|---|
|
| 94 |
+
| | | | | | | | | | |
|
| 95 |
+
|
| 96 |
+
Keep publishable text and internal production notes in separate fields. Blog, visual and post have independent readiness states. A supplied link is not a completed destination audit.
|
| 97 |
+
|
| 98 |
+
## 11. SEO/GEO blog brief
|
| 99 |
+
|
| 100 |
+
Before a search-oriented article, fill the [positioning brief](search-positioning.md#copyable-brief). Carry its intent, URL ownership and evidence state into the topic queue. A keyword list alone is not a writing brief.
|
| 101 |
+
|
| 102 |
+
## 12. Reader-value brief
|
| 103 |
+
|
| 104 |
+
- Reader and recognisable moment:
|
| 105 |
+
- Reader language / source or explicit hypothesis:
|
| 106 |
+
- What is at stake:
|
| 107 |
+
- 2–3 angles and each reader payoff:
|
| 108 |
+
- Chosen angle and supporting detail:
|
| 109 |
+
- What the reader will understand/do afterwards:
|
| 110 |
+
- Visual contribution / appropriate CTA:
|
| 111 |
+
- Editorial weaknesses to fix / real audience validation still needed:
|
| 112 |
+
|
| 113 |
+
Complete this before the keyword/production brief; do not treat it as a requirement to interrogate the user for every post.
|
| 114 |
+
|
| 115 |
+
## 13. Retrieval and diagnosis / 取数与诊断
|
| 116 |
+
|
| 117 |
+
| Field / 字段 | Source / 来源 | Window / 时间窗 | Availability / 可取状态 | Owner / 负责人 | Deadline / 截止 | Decision affected / 判别用途 |
|
| 118 |
+
|---|---|---|---|---|---|---|
|
| 119 |
+
| | | | | | | |
|
| 120 |
+
|
| 121 |
+
Then use the hypothesis table and memo in [diagnosis](diagnosis.md).
|
| 122 |
+
|
| 123 |
+
## 14. Review memo / 复盘备忘录
|
| 124 |
+
|
| 125 |
+
- Observation and dated evidence:
|
| 126 |
+
- Audience need and content payoff:
|
| 127 |
+
- Competing explanations and counterevidence:
|
| 128 |
+
- Retrieval/test results and remaining uncertainty:
|
| 129 |
+
- Continue / change / stop / retrieve / inconclusive, with reason:
|
| 130 |
+
- Owner (or unassigned role), deadline, state-file path:
|
| 131 |
+
|
| 132 |
+
## Deliverable index
|
| 133 |
+
|
| 134 |
+
Intake → 0; channel/capacity → 1; topics → 2; evidence register → 3; calendar → 4; variant review → 5; ledger → 6; handoff → 7; weekly metrics → 8; competitor research → 9; parent/social production → 10; search brief → 11; reader value → 12; retrieval/diagnosis → 13; review memo → 14. Choose the requested subset, not all templates.
|
references/website-to-social.md
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Website to social / 官网内容与社媒联动
|
| 2 |
+
|
| 3 |
+
## Decide the order
|
| 4 |
+
|
| 5 |
+
Use the intake and competitor evidence to identify the bottleneck. Searchable buyer questions plus a useful product next step justify a durable guide or comparison page. A product whose effect must be heard or seen may need a real demonstration first. A team exploring positioning may test a supported viewpoint on social before building a long article. Search discovery and social feedback can run concurrently; do not wait for rankings to distribute a ready article.
|
| 6 |
+
|
| 7 |
+
Inventory equivalent existing URLs before creating new ones. Record reader question, intent, useful evidence, conflicting claims, CTA, freshness and available performance. Choose update/create/consolidate/defer. Missing analytics means priorities are hypotheses, not proven traffic opportunities. Product competitors and search/content competitors may differ.
|
| 8 |
+
|
| 9 |
+
## Review the parent
|
| 10 |
+
|
| 11 |
+
Owned pages are sources to inspect, not ground truth. Check dated numbers, personal experiences, customer stories, superlatives, pricing and signup conditions. Contradictory copy is a verification task, not permission to choose the more attractive promise. Do not repeat unsupported claims merely with a website citation. Preserve useful sections while correcting or excluding unsupported material within authorized scope.
|
| 12 |
+
|
| 13 |
+
Answer the buyer question directly; include meaningful examples, evidence dates, relevant product links and a next step. Use diagrams/tables when they improve comprehension. Label illustrative material. Do not invent research, keyword volumes, benchmarks or cases to fill a format. Technical eligibility, indexing, rankings, AI citations and conversions are different outcomes. Structured data is not a guarantee of AI citation. SEO/GEO is not a reason to generate repetitive question pages.
|
| 14 |
+
|
| 15 |
+
## LinkedIn adaptation
|
| 16 |
+
|
| 17 |
+
Give the post a specific scenario, supported insight, reasoning and useful takeaway. The reader should benefit without clicking. Add the relevant blog/report URL where it supplies deeper detail, rather than replacing the post with a teaser. Link placement is an account-level experiment, not a universal algorithm rule.
|
| 18 |
+
|
| 19 |
+
Choose a visual by the job: comparison table for a choice, flow diagram for responsibility, document carousel for steps, dated report excerpt for evidence, real video for behavior. Text-only remains valid for a substantive argument. Do not add decorative images to disguise thin copy. Specify visual content, source, rights, caption/alt text and readiness. A brief or storyboard is not a produced visual.
|
| 20 |
+
|
| 21 |
+
Keep reviewer instructions and uncertainty checks in a separate production note. Public copy should retain material limitations naturally, without internal phrases like 'not a reported test result' unless genuinely needed to prevent misunderstanding.
|
| 22 |
+
|
| 23 |
+
## Link and measurement gates
|
| 24 |
+
|
| 25 |
+
Before publishing a linked variant, verify the article's actual public URL, correct language, mobile readability, title/share preview and matching CTA. If unpublished, keep the link pending; never send readers to an invented slug. Native standalone variants can be ready independently. Use consistent UTM parameters for external campaign links, not internal site navigation. Record campaign/content IDs and the destination URL.
|
| 26 |
+
|
| 27 |
+
Measure three paths separately: search impressions/clicks; social impressions/outbound visits; meaningful product events. Define the activation event and denominator before interpreting it. Report request is not successful generation; lead form view is not qualified inquiry; AI citation is not conversion. Attribution depends on available instrumentation/consent. No access means unknown. Select a small pilot that fits actual writing, fact-checking, design and review hours, reducing cadence when proof production takes longer.
|
| 28 |
+
|
| 29 |
+
## Depth and cooperative verification
|
| 30 |
+
|
| 31 |
+
User-confirmed first-party cases can anchor a substantive draft. Verify specific consequential claims without dismissing the entire source. A matching local source and public page confirms consistency, not independent measurement. Keep dates/estimate definitions beside numbers; put unresolved source questions in production notes. Expand with a worked decision, alternatives, reasoning and an actionable next step. Avoid short generic seeds as final deliverables.
|
| 32 |
+
|
| 33 |
+
For search-oriented parent articles, complete the [keyword and positioning brief](search-positioning.md) before writing. The social hook may differ from the search title while preserving the article’s promise.
|
references/worked-cases.md
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Worked planning cases: Ring2 and Analook
|
| 2 |
+
|
| 3 |
+
Checked 2026-11-14 using public page text. These are planning examples, not proven growth outcomes, full product tests or customer testimonials. Website claims remain claims. No private analytics, customer data or private source documents are included. Refresh these observations before reuse; the two brands are examples, not required clients or defaults.
|
| 4 |
+
|
| 5 |
+
## Ring2: explain a voice workflow before asking for a pilot
|
| 6 |
+
|
| 7 |
+
Source: https://www.ring2.ai/blog/?lang=en . The index lists human/AI/hybrid reception, handoff, appointment checking and pilot planning articles. Its 'Reading interest' numbers explicitly are not actual views. An attempted detail-page fetch failed in this check, so article-body quality remains unreviewed here.
|
| 8 |
+
|
| 9 |
+
Provisional audience: service operators evaluating after-hours inquiries; confirm market and priority use case with the user. Competitor named by the user: Retell. Its developer-oriented 2024-02-21 launch links demonstration, trial and documentation: https://news.ycombinator.com/item?id=39453402 . Transfer the proof-to-action method; do not presume the same buyer, YC amplification or available trial.
|
| 10 |
+
|
| 11 |
+
Parent decision: inspect and deepen an existing reception guide before making another overlapping article. First asset: a human/AI/hybrid decision table, with responsibility and setup requirements. A voice recording can follow once obtained and verified. Do not imply completed bookings or tested integrations from a storyboard.
|
| 12 |
+
|
| 13 |
+
LinkedIn concept A: a specific choice about who owns the callback, supported by a labeled workflow diagram. Concept B: one real demo observation with its configuration and limits, held until real media exists. Link to the reviewed guide; eventual CTA is a requirements discussion rather than an unverified instant trial.
|
| 14 |
+
|
| 15 |
+
Short illustrative seed (not a complete LinkedIn draft), without unverified product claims:
|
| 16 |
+
|
| 17 |
+
> Before choosing an answering service, decide who owns the next action. A caller may need information, a callback or a confirmed appointment. Those outcomes require different workflows. Write down the information your team needs and who will act on it—then evaluate the service against that brief.
|
| 18 |
+
|
| 19 |
+
Production note (not post copy): append the actual reviewed guide URL and informative visual when ready. Measure qualified workflow inquiries and agreed pilot scope separately from visits. No conversion baseline was collected.
|
| 20 |
+
|
| 21 |
+
## Analook: make research output useful and inspect the path to activation
|
| 22 |
+
|
| 23 |
+
Sources: https://www.analook.com/ ; https://www.analook.com/blog/competitor-seo-analysis-guide.html ; https://www.analook.com/reports/linear.html . Homepage exposes URL input, sample-report and comparison links. These are observed entry points, not a successful report-generation test. The guide is available and already covers research-to-decision, so duplicating it is not the first recommendation.
|
| 24 |
+
|
| 25 |
+
Evidence gate: the guide includes numerical performance/causal assertions and personal narratives whose underlying evidence was not established in this review. Homepage text includes both signup-free and sign-in messaging for free reports. Verify the intended flow before using a 'no signup' CTA. Do not carry unsupported efficiency or AI-citation claims into social. No current pricing, speed or accuracy promises are endorsed here.
|
| 26 |
+
|
| 27 |
+
Provisional audience: founders/growth operators choosing which competitor page or channel to investigate. Ask for the desired comparator; Similarweb/Ahrefs are comparison links on the site, not automatically identical business models or the user's selected early-stage benchmarks.
|
| 28 |
+
|
| 29 |
+
Parent decision: update the existing competitor SEO guide with a small evidence-to-action example; review the source behind any displayed report number before quoting it. Visual: a table with observation, source date, uncertainty and next experiment. Label it as a proposed analysis worksheet if not backed by an actual verified report.
|
| 30 |
+
|
| 31 |
+
LinkedIn concept A: explain why a keyword gap needs buyer-intent and feasibility checks, paired with the worksheet. Concept B: show one verified report excerpt and explain the decision it supports; hold until selected claims and media are checked. Link to the guide or exact report relevant to that post. CTA: inspect the example, then analyze a chosen URL, without an unverified signup promise.
|
| 32 |
+
|
| 33 |
+
Short illustrative seed (not a complete LinkedIn draft):
|
| 34 |
+
|
| 35 |
+
> A keyword gap is a research lead, not a publishing queue. Before assigning an article, check whether the query matches your buyer, what the existing results answer, and what evidence you can add. Record the source and date beside the proposed next step. That makes the recommendation easier to review—and easier to change when the evidence changes.
|
| 36 |
+
|
| 37 |
+
Production note: attach the reviewed worksheet and https://www.analook.com/blog/competitor-seo-analysis-guide.html only after the destination's claim review. Measure landing visits, analysis starts, successful outputs and useful follow-on actions separately. No analysis was submitted in this check.
|
| 38 |
+
|
| 39 |
+
## Decisions this pair should produce
|
| 40 |
+
|
| 41 |
+
| Decision | Ring2 | Analook |
|
| 42 |
+
|---|---|---|
|
| 43 |
+
| First bottleneck hypothesis | Understanding workflow and delivery responsibility | Turning available evidence into a decision and trying the tool |
|
| 44 |
+
| First parent | Existing guide plus scenario illustration | Existing guide plus verified report/worksheet |
|
| 45 |
+
| Proof needed | Actual configured behavior for demo claims | Traceable report facts and verified onboarding promise |
|
| 46 |
+
| Media | Workflow diagram; later real audio/video | Evidence-to-action table; later verified report excerpt |
|
| 47 |
+
| Next action | Requirements discussion | Relevant example then URL analysis |
|
| 48 |
+
| Avoid | Generic AI posts or a fake demo | More overlapping SEO pages or unsupported performance claims |
|
| 49 |
+
|
| 50 |
+
Neither hypothesis is a measured acquisition result. A proposed pilot could use one reviewed parent and two variants; select actual cadence only after capacity is known.
|
| 51 |
+
|
| 52 |
+
## Follow-up clarification, 2026-11-14
|
| 53 |
+
|
| 54 |
+
The owner confirms Analook content is authentic and useful. Earlier review gaps are not findings of falsity. Use the content as first-party material and assist with targeted verification. The Linear report’s June 23, 2026 figures (52,720 estimated monthly organic visits; 3,582 ranked keywords) match the public report and local HTML. This confirms publication consistency, not independently recomputed traffic. Retain the historical date and estimate definition when adapting.
|