issues Search Results · language:Dune language:Python language:Python language:Python language:JavaScript
Filter by
40.7M results
Summary
Adopt mdformat as the project s Markdown formatter — the documentation counterpart to ruff — and enforce it in CI and
pre-commit. Python/uv-native, so no Node toolchain.
Blocked on #168 (Python ...
maintenance
Active Bounty Scan Results
Scan Time: 2026-06-11 23:26 UTC
1. 🎯 Bounty Alert: 19 New Opportunityies found
- Repository: vansh-09/BountyScout
- Comments: 0
- Last Updated: 2026-06-11T23:20:30Z ...
bounty-alert
Why this matters
The Milestone-40 consuming-app recipe — pip install gaia-agent-email amd-gaia[api] , then gaia api — crashes at server
startup with core at main, before serving a single request: ModuleNotFoundError: ...
bug
dependencies
p1
pypi
This issue is created automatically to track contribution activity.
updation
Phase 2 — applies to future benchmark runs; depends on #5 (artifact-grounded scoring).
Problem
Much of the rater subjectivity the reliability machinery in #3 exists to measure can instead be removed ...
rubric-design
Phase 2 — applies to future benchmark runs; existing frozen results are unaffected.
Problem
Raters score the simulator s self-report, not the artifacts. Per AGENTS.md, a rater reads only the rubric, ...
harness
judge-validity
Problem
The bench s stated design principle is deterministic process layer, inference quarantined in subagents — but its single
most decision-relevant output violates it: runner.py has no compare subcommand ...
harness
Problem
Every conclusion the bench draws flows through LLM raters, and their reliability has never been measured — it is named
as a limitation in summary.md but not quantified. Without a reliability number, ...
judge-validity

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.