General-purpose coding policy for Baruch's AI agents
90
91%
Does it follow best practices?
Impact
90%
1.30xAverage score across 18 eval scenarios
Advisory
Suggest reviewing before use
**Author-Model:** line near the top of the PR body — **Author-Model:** claude-opus-4-7Co-authored-by: git trailer on any commit in the PR. Claude Code emits the trailer by default**Author-Model:** human**Author-Model:** human claude-opus-4-7Author-Model: wins when both exist (explicit beats implicit)Claude Opus 4.7 → claude-opus-4-7); unknown display names are accepted as ad-hoc model IDsREQUEST_CHANGES before reading the diffclaude-* → anthropicgpt-*, codex-* → openaigemini-* → google**Author-Model:** human claude-opus-4-7 is claude-authored for family-mismatch purposes — the openai-family reviewer runs, the claude-family reviewer skipsgpt-5.4 claude-opus-4-7): both run as degraded fallback. Avoid co-authoring across both paired families to preserve cross-family reviewgemini-2.5, custom model, human-only): both paired reviewers run — both are cross-family relative to the author; the duplicate review is accepted noise.tessl-plugin
evals
scenario-1
scenario-2
scenario-3
scenario-4
scenario-5
scenario-6
scenario-7
scenario-8
scenario-9
scenario-10
scenario-11
scenario-12
scenario-13
scenario-14
scenario-15
scenario-16
scenario-17
scenario-18
rules
skills
adopt-fork-pr
eval-curation
install-reviewer