Skip to content

P1 review epic#3

Merged
christopherarter merged 2 commits into
mainfrom
p1-review-epic
Apr 17, 2026
Merged

P1 review epic#3
christopherarter merged 2 commits into
mainfrom
p1-review-epic

Conversation

@christopherarter

Copy link
Copy Markdown
Contributor

No description provided.

Chris Arter and others added 2 commits April 17, 2026 17:42
When ruff isn't on PATH (e.g. user hasn't sourced their venv before
launching Claude Code), the rule previously returned exit 127 and
blocked edits. Use `command -v ruff || exit 0` to pass with a stderr
hint when ruff is missing, only actually running ruff when available.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
`bully bench --full` makes one real Anthropic messages.create per
fixture, capturing actual output tokens and a per-fixture cost_usd
estimate. Output tokens and USD cost appear in both per-fixture and
aggregate history records, plus the human summary table. Falls back to
count_tokens on any API failure.

Real model spend -- use for calibration runs, not every bench run.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@christopherarter
christopherarter merged commit 5326ca4 into main Apr 17, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant