Building an Automated Technical Blog System for Multi-Site Documentation

```html

What Was Done

Created a comprehensive automated technical blogging infrastructure that captures granular development work across four distinct web properties: queenofsandiego.com, sailjada.com, dangerouscentaur.com, and burialsatseasandiego.com. The system automatically generates technical blog posts at tech.[domain].com for each property, publishing detailed session-level documentation immediately after development work concludes.

Architecture Overview

The solution consists of three primary components:

Blog Generator (/Users/cb/Documents/repos/tools/tech_blog_generator.py) — Parses Claude Code session transcripts and converts them into granular, detailed technical blog posts
Infrastructure Initialization (/Users/cb/Documents/repos/tools/tech_blog_init.py) — Provisions S3 buckets, CloudFront distributions, ACM certificates, and DNS records for each tech blog domain
Session Capture Hook (/Users/cb/.claude/hooks/tech_blog_stop.sh) — Executes automatically when a Claude Code session ends, triggering blog generation and deployment

Infrastructure Setup Details

S3 and CloudFront Provisioning

For each domain, the system creates dedicated S3 buckets following the pattern tech-[domain-slug]-blog (e.g., tech-queenofsandiego-blog, tech-sailjada-blog). These buckets are configured as static website hosts with the following settings:

Block all public access disabled for CloudFront origin access
Versioning enabled for post history and rollback capability
Index document set to index.html
CloudFront Origin Access Control (OAC) for secure bucket access

CloudFront distributions are created with the following configuration:

Origin: S3 bucket with OAC authentication
Default root object: index.html
Viewer protocol policy: HTTPS redirect
Cache behaviors: 24-hour TTL for HTML, 1-year for static assets
Compression: gzip enabled for text-based content

Certificate and DNS Strategy

The infrastructure leverages existing wildcard SSL certificates:

*.queenofsandiego.com — AWS Certificate Manager (ACM) wildcard cert covers tech.queenofsandiego.com
*.sailjada.com — ACM wildcard cert covers tech.sailjada.com and tech.burialsatseasandiego.com (sailjada.com is the parent domain)
*.dangerouscentaur.com — Wildcard cert via existing CloudFront distribution (E2Q4UU71SRNTMB) using dc-sites S3 bucket as foundation

DNS records are created based on each domain's hosting provider:

Route53-hosted domains (queenofsandiego.com, sailjada.com): A-record aliases to CloudFront distributions
Namecheap-hosted domain (dangerouscentaur.com): CNAME record pointing to CloudFront domain
GoDaddy-hosted domain (burialsatseasandiego.com): CNAME record added via GoDaddy API

Blog Generator Implementation

The generator reads Claude Code session transcripts in JSONL format from ~/.claude/sessions/ and performs the following operations:

Session Parsing — Extracts tool use entries, file modifications, and command executions from the transcript
Content Generation — Constructs HTML blog posts with structured sections:
- What Was Done (executive summary)
- Technical Details (granular file changes and operations)
- Infrastructure Changes (exact resource names and ARNs)
- Key Decisions (architecture rationale)
- What's Next (follow-up work)
Credential Filtering — Removes or redacts AWS credentials, API keys, passwords, and sensitive personal data before publication
S3 Deployment — Uploads generated post HTML and updates the index with newest entries
CloudFront Invalidation — Invalidates /* cache pattern to ensure immediate visibility

Session Hook Integration

The stop hook (/Users/cb/.claude/hooks/tech_blog_stop.sh) is registered in Claude Code settings at ~/.claude/settings.json with the configuration:

"hooks": {
  "stop": {
    "script": "/Users/cb/.claude/hooks/tech_blog_stop.sh",
    "log_dir": "/Users/cb/.claude/logs",
    "timeout_seconds": 300
  }
}

The hook executes immediately after session termination, allowing developers to receive instant feedback on what was documented. A logs directory at /Users/cb/.claude/logs/ captures execution output for debugging.

Navigation Integration

The "Ship's Papers" menu on queenofsandiego.com has been updated to include a "Development Log" link pointing to tech.queenofsandiego.com. This is implemented as a dropdown item in the main navigation structure, making technical documentation easily discoverable for stakeholders like Sergio who want visibility into granular development work.

Key Technical Decisions

Wildcard Certificates — Using existing wildcard certs reduces provisioning time and eliminates DNS validation delays. This required careful domain hierarchy planning (burialsatseasandiego.com uses sailjada.com's wildcard).
Multi-Provider DNS — Supporting Route53, Namecheap, and GoDaddy required conditional logic in the infrastructure script to detect each domain's nameservers and use the appropriate provisioning API.
Granular Session Capture — Rather than high-level summaries, the system publishes all file modifications, command executions, and tool use. This enables detailed forensic analysis while filtering removes sensitive data.
CloudFront OAC — Using Origin Access Control instead of legacy OAI provides better security and audit capabilities through IAM policies.
Automatic Deployment Timing — The stop hook triggers immediately after session end, ensuring documentation stays in sync with actual code changes without manual intervention.

What's Next

Immediate follow-up items include:

Testing blog generation across all four domains with production session data
Verifying CloudFront distribution DNS propagation for all four tech blogs
Adding HTTP access tests to confirm public visibility
Creating index pages that aggregate recent posts by category
Implementing search functionality across post archives
Setting up monitoring for hook execution failures and S3 upload errors