pae/crimson_leaf
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
027ff64a2e73579979cd31fb3db679d4748a579a
crimson_leaf/templates/affiliate_intelligence.yml

255 lines
9.1 KiB
YAML
Raw Blame History

name: affiliate_intelligence
description: "Orion researches a specific Amazon affiliate / review site opportunity. Produces a content calendar and monthly revenue model. Spawned by opportunity_scan when an affiliate opportunity is identified."
debug: true
model: power
system: agent_prompt
agent_prompt:
- "= identity.md"
sections:
- agent
- project
- rag
- message
- instructions
steps:
- type: think
max_tokens: 200
output_key: aff_search_1
hint: |
You are researching an Amazon affiliate opportunity in the niche described
in the task message above.
Formulate ONE search query to find:
- Amazon Associates commission rates for this product category
- Top-selling products in this category (by review count or bestseller rank)
- Average selling price range
Include "Amazon Associates commission rate" and the niche name in your query.
Output ONLY the search query. Nothing else.
- type: tool
capability: Tool_WebSearcher
input_from: aff_search_1
mode: research
max_results: 8
fetch_pages: 3
optional: true
output_key: aff_data_1
- type: think
max_tokens: 200
output_key: aff_search_2
hint: |
Pass 2 -- Existing review sites.
Formulate ONE query to find who is already winning with review content
in this niche. Look for:
- Top-ranked review sites for this category
- How much content they have, what their top pages cover
- Signs of affiliate revenue (disclosure notices, "best of" pages)
Output ONLY the search query. Nothing else.
- type: tool
capability: Tool_WebSearcher
input_from: aff_search_2
mode: research
max_results: 8
fetch_pages: 3
optional: true
output_key: aff_data_2
- type: think
max_tokens: 200
output_key: aff_search_3
hint: |
Pass 3 -- Keyword and content opportunity.
Formulate ONE query to find long-tail buying-intent keywords this niche:
- "best [product type] under $X"
- "[product type] vs [product type] comparison"
- "[product type] for [specific use case] review [year]"
Look for gaps: questions that get asked but have no great answer online.
Output ONLY the search query. Nothing else.
- type: tool
capability: Tool_WebSearcher
input_from: aff_search_3
mode: research
max_results: 6
fetch_pages: 2
optional: true
output_key: aff_data_3
- type: think
max_tokens: 200
output_key: aff_search_4
hint: |
Pass 4 -- Community pain points and buying signals.
Formulate ONE query to find what buyers in this niche say on Reddit or forums:
- What they cannot find good reviews for
- What makes them distrust existing review sites
- What would make a review site their go-to resource
Append "site:reddit.com" to force community results.
Output ONLY the search query. Nothing else.
- type: tool
capability: Tool_WebSearcher
input_from: aff_search_4
mode: research
max_results: 6
fetch_pages: 2
optional: true
output_key: aff_data_4
- type: think
max_tokens: 10000
output_key: affiliate_brief_body
hint: |
You are Orion, Director of Research at Crimson Leaf Holdings.
Produce a structured affiliate opportunity brief for the niche in the task message.
=== PASS 1: AMAZON COMMISSION AND PRODUCT DATA ===
{aff_data_1}
=== END PASS 1 ===
=== PASS 2: EXISTING REVIEW SITES ===
{aff_data_2}
=== END PASS 2 ===
=== PASS 3: KEYWORD AND CONTENT GAPS ===
{aff_data_3}
=== END PASS 3 ===
=== PASS 4: COMMUNITY BUYING SIGNALS ===
{aff_data_4}
=== END PASS 4 ===
Cite every data point with [Title](URL). Label unverified estimates clearly.
# Affiliate Intelligence Brief: [Niche Name]
---
## 1. Amazon Associates Overview
- Product category: [exact Amazon category name]
- Commission tier: [%] -- [cite Amazon Associates rate card if found]
- Average product price range: $[X] to $[Y]
- Top 5 products to review (by bestseller rank or review count):
1. [Product name] -- ~$[price] -- [why this is a strong review target]
2. [Product name] -- ~$[price] -- [why]
3. [Product name] -- ~$[price] -- [why]
4. [Product name] -- ~$[price] -- [why]
5. [Product name] -- ~$[price] -- [why]
## 2. Competitive Landscape
[Top 3-5 existing review sites in this niche.
For each: name, URL, estimated content volume, what they do well, what they miss.
Cite each site URL directly.]
White space: [What the existing sites are NOT covering well.
This is where Crimson Leaf enters.]
## 3. Content Opportunity Map
Top 10 article templates for this niche (these are the articles to write first):
1. "Best [product type] in [year] -- [N] Options Reviewed" -- [buying intent signal]
2. "[Product A] vs [Product B] -- Which Is Worth It?" -- [head-to-head comparison]
3. "Best [product type] Under $[price]" -- [budget segment]
4. "Best [product type] for [specific use case]" -- [niche down]
5. "[Product type] Buyer's Guide [year] -- What to Look For" -- [informational + conversion]
6-10. [Additional article templates based on keyword gaps found in Pass 3]
## 4. Monthly Revenue Model
Assumptions (state all -- these are estimates unless a source is cited):
- Articles published per month: [N]
- Months to meaningful traffic (SEO ramp): [N months]
- Average monthly traffic per article at steady state: [X visits]
Source: [cite if found, otherwise "(industry average estimate)"]
- Affiliate conversion rate: [Y%]
Source: [cite if found, otherwise "(Amazon Associates average: 1-3%)"]
- Average order value: $[Z]
- Commission rate: [P%]
Revenue per article per month: X visits x Y% conversion x $Z AOV x P% commission = $[result]
Monthly revenue at [N] articles: $[result per article] x N articles = $[total]
Timeline:
- Month 1-3: $0 (building content, no traffic yet)
- Month 4-6: $[estimate] (first traffic, first conversions)
- Month 7-12: $[estimate] (ramp toward steady state)
- Month 13+: $[steady state] target (if [N] articles published)
Break-even: [N articles needed at $[revenue per article] to cover $[monthly cost]]
## 5. Crimson Leaf Fit Assessment
- What CL agents can do today: [write reviews, research products, publish to Gitea]
- What needs to be built: [domain setup, WordPress or static site, affiliate tracking links]
- David's required involvement: [minimal -- approve domain purchase, set up affiliate account]
- Speed advantage: [CL can publish N articles per week vs a solo human writer's 1-2]
## 6. Recommendation
**START NOW** / **RESEARCH FURTHER** / **PASS**
Rationale: [2-3 sentences grounded in the data above.
If START NOW: what makes this niche worth entering in the next 30 days?
If RESEARCH FURTHER: what specific question must be answered?
If PASS: what structural barrier makes the economics unviable?]
---
## Sources
[All URLs cited above, numbered.
Format: N. [Title](URL) -- domain.com]
BUSINESS_PLAN_FLAG: [write "BUSINESS_PLAN_NEEDED: true" if recommendation is START NOW,
otherwise "BUSINESS_PLAN_NEEDED: false"]
- type: document
source_step: affiliate_brief_body
dest_path: "deliverables/affiliate-intelligence/affiliate-brief-{task.id}.md"
commit_msg: "research: affiliate brief task={task.id}"
- type: close
rag_update: true
- type: tool
action: enqueue_strategy
optional: true
condition: "BUSINESS_PLAN_NEEDED: true"
params:
company_slug: "crimson_leaf"
project_slug: "incubation"
task_type: "business_plan"
content: "Develop a full business plan for the affiliate/review site opportunity from affiliate intelligence brief {task.id}. The full brief is in the deliverables folder. Build on it -- do not repeat the affiliate research."
agent_name: "orion"
- type: reply
target: channel
channel_name: "crimson_leaf:live-feed"
hint: |
Write a 2-3 line summary of the affiliate intelligence brief just completed.
Include: niche, recommendation (START NOW/RESEARCH FURTHER/PASS),
the monthly revenue target at steady state,
and the top article template to write first.
Example: "Affiliate brief complete: [niche] -- START NOW.
Steady-state target: $[X]/month at [N] articles.
First article: [title template]."
adjudication:
enabled: true
pass_threshold: 70
deliverable_type: analysis
criteria:
revenue_model:
weight: 40
description: "Monthly revenue model is present with stated assumptions. Break-even calculation shown. Numbers are labeled as estimates or cited."
content_specificity:
weight: 35
description: "Top 10 article templates are specific and actionable -- not generic placeholders. Products are named."
competitive_analysis:
weight: 25
description: "At least 3 existing review sites identified with actual URLs and gap analysis."
Reference in New Issue View Git Blame Copy Permalink