gh-aw v0.90.0 (2026-09-28) renders friction cost in step summaries, allows two noop safe-outputs by default, adds five trajectory graders with hardened generator contracts, and bumps gh-aw-mcpg to v0.4.27 / gh-aw-firewall to v0.28.27—shifting from 0.89.22’s sandbox focus toward observability and grading.

Key Takeaways

  • ✓Release v0.90.0 after 0.89.22
  • ✓Step summaries show friction cost (Cost Management)
  • ✓Default max two noop safe-outputs—skip the AI engine with zero AI Credits when work is empty
  • ✓Five new trajectory graders + hardened generator contracts
  • ✓Infra: gh-aw-mcpg v0.4.27, gh-aw-firewall v0.28.27; upload-assets summary formatting fix
🧭

Finished reading? Explore benchmark rankings & pricing

Real-world SWE-bench scores & $20/mo vs API cost break-even calculator

🏆 Leaderboard→💰 Compare Plans↗
🔗

Project Links & Resources

Direct Access
Direct access to official project resources and documentation
🐙
gh-aw v0.90.0 release
github.com/github/gh-aw/releases/tag/v0.90.0
↗
📖
Cost Management (friction / noop)
github.github.com/gh-aw/reference/cost-management/
↗
📖
Graders Specification
github.github.com/gh-aw/specs/graders-specification/
↗
📖
gh-aw Quick Start
github.github.com/gh-aw/setup/quick-start/
↗
📖
PyPI ghaw
pypi.org/project/ghaw/
↗
🐙
gh-aw
github.com/github/gh-aw
↗
🐙
#64031
github.com/github/gh-aw/pull/64031
↗
🐙
experimental Graders
github.github.com/gh-aw/experimental/trace-graders/
↗
🐙
Overview
github.github.com/gh-aw/introduction/overview/
↗
🐙
llms.txt
github.github.com/gh-aw/llms.txt
↗
🐙
0.89.22
github.com/github/gh-aw/releases/tag/v0.89.22
↗
🔬

In-Depth Technical Analysis

Background After 0.89.x sandbox hardening, gh-aw users still lacked step-level effort signals and cheap empty-run exits when there was no work for the agent. ### What shipped v0.90.0 adds friction cost in step summaries (Cost Management), default dual noop safe-outputs (#64031), five trajectory graders with hardened generators, and MCPG/firewall bumps (v0.4.27 / v0.28.27). ### Benchmarks No aggregate public scores in the release notes; measure Credits on empty runs and grader catch-rate on your workflows. ### Get started Upgrade to v0.90.0, recompile workflows, use noop to skip idle AI engines, enable new graders for trajectory gates. Docs: Quick Start, llms.txt.