moooon
Motir
Vibe the project. Bring an idea — Motir's three AI layers plan it, track it, and ship it, end to end. You're looking at Motir, built in Motir.
- Vibe Project
- Open Source
- AI Agent
- AI Loop
- 5
- suggestions
- 0
- votes
- 365
- planifiés
- 4 052
- livrés
- Reading a binary file from a connected repository returns `found` with mojibake — the provider read decodes every blob as UTF-8 and has no `binary` outcome
- Relay chat activity test races the list frame's echo against the clock
- E2E sign-in helper fails strict mode when a hidden copy of the email box is left in the DOM
- 3D / Immersive: the COLLAPSED rail draws a vertical scrollbar at laptop heights where the expanded rail and every other style fit
- (motir-core) acceptance-agent-billing flakes at a half-hour tick: the scheduled storage charge charges first, so the spec reads charged 0
- motir-core: clear open GitHub security alerts (113 Dependabot + 17 CodeQL)
- Design the /tokens page's font-set section (app/tokens/page.tsx): each locale's set with its members and default in its own glyphs, a region-forms panel, a member applied by name, and every set's faces loading only once the section is on screenMotir in eleven languages — app.motir.co and motir.co in en·zh·ja·ko·de·fr·es·it·nl·pl·pt, per-language font sets, and work content in an org content language with per-user translation
- MCP test harness never settles a request whose client hung up before the route answered — the in-flight probe fails the test after 10s
- Load every Type pairing's Latin face with the latin-ext subset in app/fonts.ts, so ą ę ł ś ź ż ó and the de·fr·es·it·nl·pt accents render in the pairing's own face (motir-core)Motir in eleven languages — app.motir.co and motir.co in en·zh·ja·ko·de·fr·es·it·nl·pl·pt, per-language font sets, and work content in an org content language with per-user translation
- Every planning session on a DeepSeek model fails on its second tool turn — DeepSeek thinks by default and returns a 400 unless each turn's `reasoning_content` is sent back, and motir-ai drops it
- Monitoring acceptance spec reads Connected instead of Degraded when the scheduled monitor poll ticks during its degraded chapter
- /device wide confirm: Approve and Deny end at 702px, below the 648px fold, with the ten-permission CLI grant
- plan.py Walk.walk crashes on a ref-less `entry: 'lay'` target: `desc` is None at `self.lay(desc + slice_, …)`
- acceptance-search-spend times out at 90s intermittently in the acceptance lane and ejects PRs from the merge queue
- (motir-core · design) TAG a PAGE in a work item's Description and Explanation — the `@` picker's Pages section, the page chip in edit and read with its unavailable state, and the Pages section's "Tagged in" labels (delta mock)Pages and work items link BOTH ways — mention or embed a work item in a page, and see on every work item the pages that mention it
- Motir's public web presence — one host per audience, and an open-core line through motir-core (motir.co · app.motir.co)
- `TokensSpecimen` mounts its own global `ThemeProvider` — dropping the specimen into a page re-stamps `<html>` with stored or app-default appearance
- A guide turn still lands on a card archived while the turn ran — tick, close (walks the archived card to Done) and file_bug all report `landed`
- (motir-core) `validate_plan` and projected `validate_work_item` report EVERY bad edge in the committed tree, so any old uncovered or cross-level edge makes every plan on the project read `valid: false` with no blocker and no rejection — narrow the edge findings to the ones the plan introduces
- app.motir.co/ 308s to motir.co instead of landing a signed-in reader on /workbench and a signed-out one on /sign-in