A pattern I keep seeing in AI is individuals and teams (engineering orgs I've worked with and also vibe coders) publishing their development workflow. There's usually a few glaring problems though:
1. It's a human not ai workflow - "this is how I would solve the task"
2. It's domain specific - "for this given problem, solve it this particular way"
3. It's dated the moment you publish it!
That's coming from someone who also has their workflow (a set of loosely coupled skills with routing calls).
If teams and individuals aren't using AI to continually improve their own processes, skills, and workflows, you've lost the plot. There should not be a single one size fits all approach.
What are your thoughts? I could be completely off base so I'm curious to know others opinions
Issue and context tracking is the biggest problem developing software in 2026 (for humans and AI agents). The vast majority of teams use @github yet I can't customize what shows up or the graph of information associated with PRs natively. It should be bring your own.