We document a Continue.dev setup in docs/setup.md § Continue.dev, but the maintainer doesn't use Continue.dev day-to-day — this needs field validation.
If you use Continue.dev, try LID on a real task and post here:
- Did it work? Did the agent actually follow the workflow — walk the phases, stop for your review, write EARS specs and
@spec annotations?
- What did you have to do to set it up, beyond what docs/setup.md says?
- Any evidence you can share — transcript snippets, a repo link, screenshots.
Continue's adapter inlines the methodology (it doesn't auto-read AGENTS.md), so drift between the rule file and AGENTS.md is a specific thing to watch for.
Negative reports are just as valuable as positive ones — "the agent ignored the rule file entirely" is exactly what we need to know. Fixes to the setup.md entry are welcome as PRs (trivial-change tier per CONTRIBUTING.md — no arrow walk needed).
We document a Continue.dev setup in docs/setup.md § Continue.dev, but the maintainer doesn't use Continue.dev day-to-day — this needs field validation.
If you use Continue.dev, try LID on a real task and post here:
@specannotations?Continue's adapter inlines the methodology (it doesn't auto-read
AGENTS.md), so drift between the rule file andAGENTS.mdis a specific thing to watch for.Negative reports are just as valuable as positive ones — "the agent ignored the rule file entirely" is exactly what we need to know. Fixes to the setup.md entry are welcome as PRs (trivial-change tier per CONTRIBUTING.md — no arrow walk needed).