Skip to content

fix(setup): default max agent turns to 90 (#4018) - #4031

Open
avishj wants to merge 2 commits into
NousResearch:mainfrom
avishj:fix/max-iterations
Open

fix(setup): default max agent turns to 90 (#4018)#4031
avishj wants to merge 2 commits into
NousResearch:mainfrom
avishj:fix/max-iterations

Conversation

@avishj

@avishj avishj commented Mar 30, 2026

Copy link
Copy Markdown

What does this PR do?

As described in #4018, there is an inconsistency between 60/90 for max agent turns. I've fixed it to be 90 consistently everywhere it was still at 60.

Related Issue

Fixes #4018

Type of Change

  • 🐛 Bug fix (non-breaking change that fixes an issue)
  • ✨ New feature (non-breaking change that adds functionality)
  • 🔒 Security fix
  • 📝 Documentation update
  • ✅ Tests (adding or improving test coverage)
  • ♻️ Refactor (no behavior change)
  • 🎯 New skill (bundled or hub)

Changes Made

  • cli-config.yaml.example: max_turns 60 -> 90
  • cli.py: docstring default 60 -> 90
  • environments/benchmarks/tblite/default.yaml: max_agent_turns 60 -> 90
  • environments/benchmarks/tblite/local.yaml: max_agent_turns 60 -> 90
  • environments/benchmarks/tblite/local_vllm.yaml: max_agent_turns 60 -> 90
  • environments/benchmarks/terminalbench_2/default.yaml: max_agent_turns 60 -> 90
  • website/docs/developer-guide/environments.md: Updated documentation examples for max_agent_turns default to 90

How to Test

  1. Install hermes via setup-hermes.sh.
  2. Prompt shows 90 in both locations.

Checklist

Code

  • I've read the Contributing Guide
  • My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
  • I searched for existing PRs to make sure this isn't a duplicate (existing PR is incorrect)
  • My PR contains only changes related to this fix/feature (no unrelated commits)
  • I've tested on my platform: macOS 15.7.3

Documentation & Housekeeping

  • I've updated relevant documentation (README, docs/, docstrings)
  • I've updated cli-config.yaml.example if I added/changed config keys

Screenshots / Logs

Screenshot 2026-03-30 at 16 24 17

@avishj avishj changed the title fix: #4018 fix default max agent turns to 90 fix(setup): default max agent turns to 90 (#4018) Mar 30, 2026
@alt-glitch alt-glitch added type/docs Documentation improvements P3 Low — cosmetic, nice to have comp/cli CLI entry point, hermes_cli/, setup wizard duplicate This issue or pull request already exists labels May 2, 2026
@alt-glitch

Copy link
Copy Markdown
Collaborator

Likely duplicate of #4020 — both fix the same max_turns 60/90 inconsistency from #4018.

1 similar comment
@alt-glitch

Copy link
Copy Markdown
Collaborator

Likely duplicate of #4020 — both fix the same max_turns 60/90 inconsistency from #4018.

@teknium1

Copy link
Copy Markdown
Contributor

Thanks for tracing the inconsistent max-turn documentation.

Problems

  • This branch is now conflicting: its four environments/benchmarks/... YAML targets and website/docs/developer-guide/environments.md are absent on current main. The remaining two corrections are still applicable: cli-config.yaml.example:684 and cli.py:15795 still say 60, while hermes_cli/config.py:991 defines the runtime default as 90.

Suggested changes

  • Salvage the two live documentation/example edits and drop the removed-file hunks. Current CLI coverage already verifies the runtime default is 90 in tests/cli/test_cli_init.py:60-72.

Automated hermes-sweeper review.

@teknium1 teknium1 added the sweeper:blast-contained Sweeper blast radius: contained — one narrow path / opt-in / few users label Jul 12, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

comp/cli CLI entry point, hermes_cli/, setup wizard duplicate This issue or pull request already exists P3 Low — cosmetic, nice to have sweeper:blast-contained Sweeper blast radius: contained — one narrow path / opt-in / few users type/docs Documentation improvements

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[Setup]: Max Tool Calls is Inconsistent (60/90)

3 participants