Open it like you'd open a video.
Four synchronized views. File City shows where. Sequence shows how. Code shows what. Plain English explains why. Click one, the others follow. Click a line, leave a note, right where it belongs.
Make a trail when you need: clarity
Trails are a byproduct of working, not extra work. See an example trail.
When the whole team understands it,
the whole team owns it.
Agents made one person fast. Code Trails make the whole team keep pace. One link at a time, understanding stops living in one head and starts compounding into something everyone can find, follow, and build on.
Everyone sees what got built.
Why it matters, where it lives, how it works. The context that makes code make sense, available to the whole team.
Understanding that compounds.
Every trail makes the next one easier. One link at a time, your team gets smarter about its own code.
Code everyone can stand behind.
When understanding is shared, confidence is too. Decisions stop living in one head and become something the whole team owns.
Make your own Trail
Drop this prompt into your agent.
Read the author-local-investigation-trail skill at https://github.com/principal-ai/skills/blob/main/author-local-investigation-trail/SKILL.md and use it to walk me through <topic> locally as an investigation trail.Built by engineers who got tired of clone-the-repo to answer one question.