"I followed the guide perfectly. Built all seven agents. Wrote the orchestration file. Then I opened a different coding agent and... nothing. It ignored everything."
If that sounds familiar, the pipeline isn't broken. Your agent just never read it.
Agents Don't Agree on Much
Every coding agent reads instructions, but they disagree on the details that matter here. Three differences break a pipeline:
- •Which instructions file loads on its own. Some read
AGENTS.mdat the start of every session. Others only read a file with their own name, and some read nothing until you tell them to. - •Whether you can define commands. Some let you save "run the pipeline" as a slash command or workflow file. Others make you type the request every time.
- •Whether separate agents get their own files. Some let each stage live in its own agent definition (sometimes with its own model). Others run everything as one agent reading one set of instructions.
And this changes. Tools rename their files and add features every few months, which is why this page names none of them.
What It Looks Like When the File Didn't Load
- ✗No stops: the agent runs from classify to publish prep in one go
- ✗Answered for you: it fills in the focus questions itself, which is where invented stories come from
- ✗Voice drift: your voice guide and banned words never got read
- ✗Files in the wrong place: no numbered files in
drafts/[slug]/, so nothing survives the session
Check Your Agent
Fill this in for the tool you use. Tick what it supports. Every box you leave empty tells you which shim you need in Part 3. (Check the docs. Don't take the agent's word for its own features.)
Ask your agent what it supports
Try itPaste into the coding agent you want to use. Then check its answers against the docs.
I'm setting up a writing pipeline whose rules live in AGENTS.md and agents/*.md in this folder. Before I change anything, answer these about yourself:
1. Which instructions file names do you load automatically at the start of a session? Do you read AGENTS.md?
2. Can I define my own commands (slash commands or workflow files)? If so, where do they live and what format do they need?
3. Can I define separate agents, each with its own instructions file? If so, where, and can each one set its own model?
4. Do you read files from subfolders like agents/ on your own, or only when told to?
5. Can you stop mid-task and wait for my reply, or do you tend to finish the whole job in one go?
Point me to the documentation for each answer. If you aren't sure, say "not sure". Don't guess.