iris-ares/PR-Lens

Business-context-aware Verify Checklist for your PRs. Pick a provider by name.

Trust Signals

name	description	required	default
llm-provider	LLM provider to use. One of: openai, anthropic, deepseek, google, groq, mistral, xai, openrouter, openai-compatible. Use "openai-compatible" together with llm-base-url for self-hosted vLLM or custom endpoints.	yes	—
llm-api-key	API key for the selected provider. Will be masked via ::add-mask::.	yes	—
llm-model	Provider-native model identifier. Examples: "gpt-4.1" (openai), "claude-sonnet-4-5-20250929" (anthropic), "deepseek-chat" (deepseek), "openai/gpt-4o" (openrouter), "qwen2.5-7b" (openai-compatible).	yes	—
llm-base-url	Base URL for the endpoint. Required when llm-provider is "openai-compatible"; ignored otherwise.	no	—
github-token	Token used to read the PR and write the sticky comment. Defaults to the job token.	no	${{ github.token }}
memory-path	Path (repo-relative) to the business memory markdown the agent reads and proposes patches to.	no	.pr-lens/business.md
max-files	Cap on number of changed files passed to the agent.	no	50
max-diff-tokens	Soft cap (estimated tokens) on the compressed diff added to the prompt.	no	15000
step-budget	Hard cap on agent tool-call steps before finalization.	no	12
comment-marker	HTML comment used to identify the sticky comment across runs. Do not change unless you want a fresh comment.	no	<!-- pr-lens:main -->

no outputs