Writing · Tag

Articles tagged "ai-assisted-development".

9 articles on this topic.

Articles tagged "ai-assisted-development"

·12 min read · chrome-extension , manifest-v3 , chrome-web-store

Shipping a Manifest V3 Chrome extension: the gates nobody mentions

Trader verification, publisher identity, the URL slug gotcha, version-bump CI. The async gates that turn a one-day project into a three-week project.

Shipping a Manifest V3 Chrome extension: the gates nobody mentions: cover image

·9 min read · methodology , ai-assisted-development , refactoring

Phased migrations with per-phase verification gates

Big-bang rewrites of 6+ files break things you cannot predict. Phased migration with explicit gates is slower and catches the regressions early.

Phased migrations with per-phase verification gates: cover image

·10 min read · claude-code , claude-md , lovable

How I document AI-built projects: a CLAUDE.md, ISSUES.md, and prompts/ workflow

The four-part system I install on day one of every inherited Lovable, Cursor, or Claude Code project: CLAUDE.md template, ISSUES.md learning log, prompts/ folder, two-AI workflow.

How I document AI-built projects: a CLAUDE.md, ISSUES.md, and prompts/ workflow: cover image

·8 min read · supabase , supabase-auth , supabase-publishable-key

Migrating to Supabase publishable keys broke my Chrome extension. Here is the fix.

Supabase's new sb_publishable_* keys return 401 when sent as raw apikey headers, and the JS SDK defaults to localStorage which a Manifest V3 Chrome extension cannot use. Here is the migration: deleting hand-rolled fetch wrappers, switching to @supabase/supabase-js, and the chrome.storage.local adapter that keeps sessions persistent.

Migrating to Supabase publishable keys broke my Chrome extension. Here is the fix.: cover image

·9 min read · lovable , lovable-security , supabase

How to audit a Lovable app after the BOLA disclosure: a 6-hour rotation playbook

Lovable's April 2026 BOLA vulnerability exposed projects created before November 2025. Here is the audit checklist I ran on a client's Supabase-backed Lovable app, the 6-hour key rotation that followed, and the Chrome-extension SDK migration nobody warned us about.

How to audit a Lovable app after the BOLA disclosure: a 6-hour rotation playbook: cover image

·9 min read · issues-md , claude-code , methodology

ISSUES.md as a learning log: a 78-issue tracker pattern for AI-built projects

ISSUES.md is a flat markdown bug tracker that doubles as a regression detector and training doc. The pattern that works after 78 closed entries.

ISSUES.md as a learning log: a 78-issue tracker pattern for AI-built projects: cover image

·10 min read · aider , pagerank , tree-sitter

How Aider's repomap uses PageRank and tree-sitter to rank your codebase

Aider parses your repo with tree-sitter, builds a symbol graph, and runs PageRank to pick which files the LLM should see. Here is how it works.

How Aider's repomap uses PageRank and tree-sitter to rank your codebase: cover image

·8 min read · cursor , fast-apply , diffs

Why AI coding tools rewrite full files instead of using diffs

Cursor fast apply, Morph, and Aider all converge: full-file rewrites beat diffs under 400 lines. Three reasons rooted in how language models work.

Why AI coding tools rewrite full files instead of using diffs: cover image

·12 min read · vibe-coding , cursor , claude-code

How AI coding tools actually edit code: 6 truths from 4,000+ hours of building software

After 4,000+ hours of building software, the last stretch with Cursor, Claude Code, Aider, and Lovable, here is how AI coding tools actually edit code under the hood.

How AI coding tools actually edit code: 6 truths from 4,000+ hours of building software: cover image
← All articles