How to Write a Continue.dev Config from a Repo Stack Brief

IDE configs wander when a Continue.yaml grows a Cursor AGENTS.md and a model you do not subscribe to. Continue.dev is its own file. Lock the model id and refuse the neighboring assistants.
The matching generator is the Continue.dev Config from a Repo Stack Brief prompt. Browse related cards in the PromptDig library (Browse more prompts). When a filled run survives, share the version you actually use (Share a prompt).
Version-lock Continue and name the schema
Write a Continue.dev config.yaml from a repo stack brief. No invented models or other-IDE rule files. Start by filling Inputs, not by asking the model to remember last week's run. If a field is blank, write NONE or NOT IN INPUTS and leave it blank through Generate. The card is built so the model cannot honestly invent a number, owner, URL, or command that you did not paste.
Paste these fields before you hit run:
Continue version or config schema I want: [Version]
Repo languages, package managers, test command: [Stack]
Models I may call (provider plus model id): [Models]
Context providers I allow: [Context]
Rules I want in config (never-list, style): [Rules]
MCP or tools I actually have: [Tools or NONE]
Auth note (env var name only): [Auth]
What other assistant files I must not write: [NotFiles]
Tab autocomplete model or NONE: [Tab]
Words I must not use: [Banned]
That inventory is the honesty ledger. Anything that does not appear there is forbidden in the draft. If you catch yourself adding a nice-to-have after the run, you are no longer using the card. You are ghostwriting. Put the extra fact in Inputs and run again.
Models only from the list you pasted
Generate is numbered on purpose. Do not skip a step because the first paragraph looked done. The early steps exist to stop later prose from smuggling claims.
Walk the Generate list in order:
- Version lock: quote Version. If unknown, write config.yaml for Continue 1.1.x style and mark schema unknown.
- Honesty ledger: languages, test command, model ids, providers, context providers, tools. Forbidden: models not in Models, files in NotFiles.
- config.yaml sketch: models list from Models only, tabAutocompleteModel from Tab or omit, contextProviders from Context.
- Rules block: never-list from Rules. Include Stack test command as the only test command.
- Tools: MCP only if Tools names them. If NONE, omit.
- Auth: print only the env var name from Auth. Do not print a token.
- Refuse other files: list NotFiles you will not emit.
- Compliance pass: quote Banned words, invented model ids. Cut them.
If a step asks for a version lock, quote the version from Inputs in the output. If a step asks for a refuse list, keep the refuse list in the published artifact, not in a sidebar you delete. Reviewers should see what the model was not allowed to do.
Omit tab autocomplete when Tab is NONE
Most failures are the same shape: a missing field gets a confident fill. A conversion rate appears. A Gradle task appears. A flash point appears. A caption appears on a job that asked for slide text only. Your review is to search the draft for numbers, names, and commands, then grep Inputs. No match means cut.
Honor the constraints as hard stops, not vibes:
- Continue.dev config.yaml only. Do not write Cursor, Copilot, Cline, Windsurf, or JetBrains AI files.
- Never invent a model id.
- Test command only from Stack.
- If Tab is NONE, omit autocomplete rather than guessing a small model.
- No emojis.
When the card says not legal advice, not certification, not an exam dump, or not a caption engine, that sentence belongs at the top of the output. Deleting it to look more finished is how you inherit risk.
Do not emit AGENTS.md, Cline, or Copilot files
Finish with the compliance pass the prompt already asks for. Quote the banned-word hits. Cut them. Print character counts when the job has a cap. Print word counts when the job has a budget. List gaps as gaps. Five missing facts are more useful than one smooth paragraph.
Tags on the card (continue.dev config yaml, continue config from repo, ide assistant continue) are a reminder of the job shape, not an invitation to wander into a neighboring cluster. If you need a different surface, open a different PromptDig card rather than stretching this one.
Fill the card, then run
Replace every bracket. Run on ChatGPT, Claude, or Gemini. Read the ledger first, then the artifact. If the model invents a commit, KPI, DOI, PEL, bid, or logo, discard the run. Tighten Inputs. Run again. Share the filled card that survived, not the first draft that sounded done.