Release v2.0.0 · yetanotherco/actions

Breaking changes

pr_code_review_claude.yml is removed. Use pr_review_claude.yml.
pr_review_claude.yml defaults model to claude-opus-5-5 instead of sonnet, at about twice the cost. Set model: sonnet to keep the old behaviour.
pr_review_kimi.yml defaults model to kimi-k3.1-pro with reasoning_effort: high. Moonshot retired the old kimi-k2.5.
pr_review_codex.yml runs gpt-6.1-sol instead of the Codex CLI's default model. Override it with the new model input.

pr_review_openrouter.yml runs DeepSeek, Grok, Gemini or any other OpenRouter model with one OPENROUTER_API_KEY. See the README for the recommended model, reasoning_effort and max_bytes per model.
custom_prompt is optional. Without it, the reviewers read .github/prompts/ai-review.md from the PR's base branch, then fall back to this repository's default prompt.
Kimi and OpenRouter reviews read the diff from git, so PRs over the pulls API's 20000-line limit are reviewed (truncated to max_lines / max_bytes, with a note in the comment).
The Claude and Codex reviewers no longer get write-capable GitHub tokens, and pr_review_claude.yml no longer needs id-token: write.

fix: pin external actions to commit SHAs and update them by @MarcosNicolau in #14
ci: review this repository's PRs with its own AI review workflows by @JuArce in #18
fix: keep write-capable GitHub tokens away from the review agents by @JuArce in #17
feat: add Kimi K3.1, DeepSeek, Grok and Gemini reviewers by @MarcosNicolau in #15
feat: read the review prompt from .github/prompts/ai-review.md by @MarcosNicolau in #16
ci: review this repository with DeepSeek, Grok and Gemini by @MarcosNicolau in #20
feat: default the reviewers to flagship models by @MarcosNicolau in #21

Full Changelog: v1.0.4...v2.0.0