issues Search Results · language:Edge language:TypeScript language:PHP language:JavaScript language:Python language:Python
Filter by
18.4M results
Performance Regression Detected
Commit: 9f9e97be Run: https://github.com/ROCm/ATOM/actions/runs/26719406873 Date: 2026-06-01T02:53:20.502928+00:00
Regressed Configurations
| Model | ISL/OSL | Conc | ...
performance
regression
발견 경위
환경 3 runtime acceptance 시도(2026-06-01) 중 발견: devtools-mcp 프로세스가 2일 17시간 동안 백그라운드에 살아있고, list_pages가 tunnel.up: true로
보고하지만 실제 wss handshake는 실패(build_attach_url 후 90s wait_for_attach 매번 timeout). ...
bug
roadmap
[Proposal] Adversarial Test Suite — 52-Scenario Security Assessment
Summary
We implemented a comprehensive adversarial test suite for AGT, covering 52 scenarios across:
- OWASP Agentic AI Top 10 (all ...
Sprint Retrospective - 4 Weeks
| Satisfaction-level and Value | Justification |
| --: | :-: |
| Very Happy (5) | 🙂 전반적으로 하고 싶었던 실험·프로젝트·휴식·운동을 모두 시도하며, 자동화/운동/라이프스타일 측면에서 균형 있게 성장한 한 달 |
| Happy (4) | ...
👩💻Monthly
Kiro Product
IDE
Feature Description
Feature Request
Kiro is built on top of VS Code, but currently it does not support the native macOS Merge All Windows functionality.
This feature is available ...
pending-triage
Summary
A solo /afk run worker has no idle/stall reaper. The passive stall detector + hard stall reaper described in the
SKILL.md exist only in Fleet Mode (the supervisor). A single run worker relies ...
priority:high
ready-for-agent
slice:afk
type:bug
Summary
The AFK inner agent can write a self-deadlocking polling loop that hangs the worker forever. Observed live on #302
(worker wKXWG): the agent ran vitest in the background, then polled with
until ...
priority:high
ready-for-agent
slice:afk
type:bug
Summary
User report: crane eval 套件這次給的訊號不可靠且互相矛盾 . Confirmed multiple deterministic bugs across the eval/audit tools. Verdicts
are invalid and some auto-fixes are destructive (manuscript corruption). ...
bug
Add analysis of how speculative/parallel sampling branches affect self-consistency.
Trend context: Parallel sampling strategies are common in 2026 inference.
Requirements
- Compare self-reports across ...
enhancement
research
scoring
Add analysis of idle/ offline cycle effects for agents that run background processing.
Trend context: Background/idle processing is appearing in 2026 agent architectures.
Requirements
- Probe state ...
enhancement
research
scoring

Learn how you can use GitHub Issues to plan and track your work.
Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub IssuesProTip! Restrict your search to the title by using the in:title qualifier.