Claude Code and Codex skills are directories, not files
Why Claude Code and Codex discover agent skills as a directory containing SKILL.md, not a flat markdown file, and what silently breaks when an installer gets that wrong.
Why Claude Code and Codex discover agent skills as a directory containing SKILL.md, not a flat markdown file, and what silently breaks when an installer gets that wrong.
How a human-triggered ‘@claude review’ workflow got locked down to read-only, bounded-diff, revision-verified reviews after the official tag-mode example turned out to be implementation-capable by default.
How the waaseyaa/agent-output package shrinks PHPUnit, PHPStan, and bin/check-* output to compact NDJSON envelopes so AI agents do not drown their context window in CI noise.
AI assistants estimate effort using training data from before they existed, then push back when you try to spec faster. Working around it eats hours every week.
Six Hugo shortcodes that give your PaperMod blog callouts, steps, pull quotes, stats, before/after comparisons, and CTAs — built in one vibe coding session.
First-run lessons from an automated content pipeline. Noise, human-only merges, and a backwards production step surfaced in 24 hours.
Stop AI from stripping your writing voice. Encode your style rules, review the output, and feed corrections back in.
Claudia turns Claude Code into a personal chief of staff that remembers your relationships, tracks your commitments, and helps you make better decisions.
Use a Claude Code startup hook to surface untriaged GitHub issues and stale milestones before you write any code.
How in-memory implementations, a layered test strategy, and AI-assisted test generation keep a 48-package PHP monorepo testable.