pullrequests Search Results · language:Edge language:TypeScript language:HTML language:HTML language:Python language:TypeScript
Filter by
159M results
Summary
setup.sh and setup.ps1 already resolve the AMD gfx target through several methods (rocminfo / hipinfo / amd-smi,
GPU-name inference, and the UNSLOTH_ROCM_GFX_ARCH override), but they never passed ...
DRAFT — do not merge yet. Adversarial Codex review returned VERDICT: BUG. Finding #3 is fixed + regression-tested;
findings #1/#2/#6 need one more iteration (below) before this is mergeable. Opening as ...
Summary
Adds a hero image to the README showing the sixteen testing techniques the skill teaches, organised into the skill s own
three-tier structure and framed by the red-green-refactor rhythm. A reader ...
Summary
Real-deploy UX gap, same exploration as PR #61: the Account screen is a modal route (deliberately hides the BottomNav so
chrome doesn t compete with modal affordances) but had no in-app way to ...
What
Adds a second internal mode for webhook-triggered conversations. Until now a definition with action= conversation had
to declare an outboundChannel + outboundTarget; the matched event always sent ...
🎯 What: Extracted the massive inline SQL schemas, index creations, and migrations out of init_db in
backend/core/database.py and into separate lists in a new backend/core/database_schema.py file. 💡 Why: ...
Summary
- add a Keep a Changelog style CHANGELOG.md for v0.1.0
- include post-v0.1.0 install-doc clarification under Unreleased
- include CHANGELOG.md in the package files list
Validation
- ...
Summary
- add a persisted setting for progress-based session title/phase refresh with a configurable turn cadence
- re-run session naming reminders on later OpenAI Codex milestones so long sessions ...
Summary
- Add a GitHub Pages deployment workflow for the Vite app.
- Build with npm run build and upload the generated dist directory.
- Deploy using GitHub Pages Actions with the github-pages environment. ...
Implements caching of data/inventory.yaml using actions/cache in case of failure, mimicking the V2 AI pipeline.