kevinpatrickrobbins/codebase-audit/content-creator

Blog Content Pipeline

A ticketing-system-style workflow for managing blog content from idea to publication, built around three interconnected Markdown documents that serve as the single source of truth for all content operations.

System Architecture

The pipeline uses three documents that act as stages in a Kanban-style workflow:

IDEAS BACKLOG ──▶ APPROVED QUEUE ──▶ PUBLISHED TRACKER
(Brainstorm)      (Greenlit)         (Completed)

Each document lives in the user's working directory and is managed through structured Markdown with consistent formatting so the agent can read, update, and move items between documents programmatically.

Document Locations

All three documents are created in the user's current working directory (or a directory they specify). The default filenames are:

1. content-ideas-backlog.md — The brainstorming and idea development document
2. content-approved-queue.md — Approved ideas ready for writing
3. content-published-tracker.md — Completed and published articles

Core Workflows

1. INITIALIZE — Setting Up the Pipeline

When the user asks to set up or start their content pipeline:

1. Read references/ideas-backlog-template.md for the Ideas Backlog structure
2. Read references/approved-queue-template.md for the Approved Queue structure
3. Read references/published-tracker-template.md for the Published Tracker structure
4. Create all three files using those templates
5. Ask the user for their blog's niche/topic area, target audience, and publishing cadence
6. Record these in the Ideas Backlog header section

2. BRAINSTORM — Generating New Ideas

When the user asks to brainstorm or generate article ideas:

1. Read references/ideation-strategies.md for the full ideation framework
2. Ask the user which ideation strategy to use (or suggest one based on context):

• Content Pillar Mapping — Build clusters around core topics
• Audience Pain Point Mining — Address real problems your readers face
• Keyword Opportunity Analysis — Target search-driven topics
• Content Gap Analysis — Find what competitors cover that you don't
• Seasonal/Trend Calendar — Align with timely events and trends
• Repurpose & Remix — Spin new angles from existing high-performers

1. Generate 5-10 ideas using the selected strategy
2. For each idea, populate the standard idea card fields (see Idea Card Format below)
3. Append new ideas to the Ideas Backlog document
4. Present the ideas to the user for initial reaction

3. REVIEW — Evaluating Ideas for Approval

When the user asks to review their backlog or approve ideas:

1. Read the current Ideas Backlog
2. Present ideas grouped by status (New, Under Review, Needs Research, Deferred)
3. For each idea under review, present a brief scorecard:

• Strategic Fit — Does it align with the blog's mission and content pillars?
• Audience Value — Will readers find this genuinely useful or interesting?
• Effort vs. Impact — Is the expected return worth the production effort?
• Timeliness — Is there urgency or a seasonal window?
• Uniqueness — Does this offer a fresh angle not already saturated?

1. Ask the user to approve, defer, or reject each reviewed idea
2. Move approved ideas to the Approved Queue with a target date
3. Update deferred/rejected ideas in the backlog with the reason

4. PRODUCE — Managing the Writing Queue

When the user asks about their writing queue or what to work on next:

1. Read the Approved Queue
2. Present articles sorted by priority and target date
3. For articles marked "In Progress," show current status and any blockers
4. Suggest which article to tackle next based on:

• Approaching deadlines
• Strategic priority
• Content clustering (publish related pieces together)
• Seasonal relevance windows

1. When an article is completed, prompt the user to move it to the Published Tracker

5. PUBLISH — Recording Completed Articles

When the user marks an article as done or asks to log a published piece:

1. Collect the publication details (URL, publish date, actual title)
2. Move the item from the Approved Queue to the Published Tracker
3. Remove it from the Approved Queue
4. Add initial performance tracking fields (to be updated later)
5. Check if the published article unlocks any related ideas in the backlog

6. REPORT — Pipeline Health Check

When the user asks for a status report or pipeline overview:

1. Read all three documents
2. Generate a summary showing:

• Ideas in backlog (by status)
• Articles in the approved queue (by stage)
• Recently published articles
• Pipeline velocity (ideas → approved → published rates)
• Content pillar coverage balance
• Upcoming deadlines or seasonal windows at risk

Idea Card Format

Every idea uses this consistent format across all three documents:

### IDEA-{NNN}: {Working Title}
 
| Field            | Value                                    |
|------------------|------------------------------------------|
| **Status**       | {New / Under Review / Approved / etc.}   |
| **Content Pillar** | {Which core topic this belongs to}     |
| **Format**       | {How-to / Listicle / Deep-dive / etc.}   |
| **Target Audience** | {Primary reader persona}              |
| **Angle/Hook**   | {What makes this piece unique}           |
| **SEO Target**   | {Primary keyword or topic cluster}       |
| **Effort**       | {Low / Medium / High}                    |
| **Priority**     | {P1-Critical / P2-High / P3-Normal / P4-Low} |
| **Created**      | {YYYY-MM-DD}                             |
| **Target Date**  | {YYYY-MM-DD or TBD}                      |
| **Notes**        | {Additional context, research links}     |

When an idea moves to the Approved Queue, these fields are added:

| **Approved Date**  | {YYYY-MM-DD}                           |
| **Outline Status** | {Not Started / Draft / Final}          |
| **Draft Status**   | {Not Started / In Progress / Review / Final} |
| **Assigned To**    | {Writer name or "Self"}                |

When an article moves to the Published Tracker, these fields are added:

| **Published Date** | {YYYY-MM-DD}                           |
| **Final Title**    | {Actual published title}               |
| **URL**            | {Link to published article}            |
| **Word Count**     | {Approximate}                          |
| **Performance**    | {Notes on traffic, engagement, etc.}   |

ID Assignment

Ideas are assigned sequential IDs in the format IDEA-001, IDEA-002, etc. The ID stays with the idea as it moves through the pipeline so you can always trace an article back to its origin. When generating new IDs, read the Ideas Backlog to find the highest existing ID and increment from there.

Best Practices Encoded in This System

These principles from professional editorial workflows are built into the pipeline:

• Systematic ideation over inspiration-dependent brainstorming — The ideation strategies

in references/ideation-strategies.md ensure a steady flow of ideas rather than waiting for inspiration to strike.

• Content pillars for topical authority — Organizing ideas by content pillar helps build

depth in key topic areas rather than publishing scattered, unrelated posts.

• Effort-aware scheduling — Matching production capacity to ambition prevents the most

common content calendar failure: over-scheduling.

• Cluster sequencing — Publishing related pieces in sequence builds internal linking

and topical authority faster than isolated posts.

• Pipeline visibility — The three-document system gives at-a-glance visibility into

every stage, making it easy to spot bottlenecks and gaps.

• Quarterly review cycles — The Published Tracker supports retrospective analysis to

inform future ideation and strategy adjustments.

Reference Files

Read these files for detailed guidance on specific workflows:

Important Behaviors

• Always read the existing documents before making changes to avoid overwriting content
• Preserve all existing ideas when appending new ones
• When moving items between documents, copy the full card (don't lose fields)
• Keep IDs consistent across all three documents
• When the user says "what should I write next?" — read the Approved Queue and give a

specific recommendation with reasoning, not just a list

• When brainstorming, generate genuinely varied ideas, not slight variations of the same topic
• Default to creating all three documents if they don't exist yet when the user first

engages with any content pipeline workflow