propstreet/Pro PR Reviewer

Uses Azure OpenAI to review diffs, then posts code review comments on GitHub PRs.

Trust Signals

Scorecard Score: not yet scored
Maintenance Recency: Stale
License: None

Inputs

name	description	required	default
azureOpenAIEndpoint	Azure OpenAI endpoint (e.g., https://your-resource.openai.azure.com/openai/deployments/...)	yes	—
azureOpenAIDeployment	Azure OpenAI deployment (e.g., your-resource)	yes	—
azureOpenAIKey	API key for Azure OpenAI	yes	—
base	Base commit SHA to compare against.	no	—
head	Head commit SHA to compare against.	no	—
severity	Minimum severity level required to post a request changes review comment (info, warning, or error).	no	error
reasoningEffort	Reasoning effort level (none, minimal, low, medium, high, or xhigh). 'none' disables reasoning (gpt-5.1+); 'minimal' is the legacy equivalent (gpt-5 only). 'xhigh' enables maximum reasoning depth on supported models. Using unsupported levels may cause API errors or increase latency/cost. For review purposes 'high' is recommended.	no	medium
tokenLimit	Maximum number of tokens to process.	no	50000
commitLimit	Maximum number of commits to load for 'entire-pr' diff mode.	no	100
exclude	Glob patterns to exclude files from review (e.g. '.test.ts' or 'dist//')	no	""
customPrompt	Custom instructions appended to the system prompt for the LLM reviewer (max 1000 chars). Example: 'Focus on security issues and best practices'	no	""
backgroundMode	Enable background mode for long-running requests (recommended for GPT-5-Pro). Options: 'enabled', 'disabled'.	no	disabled
backgroundMaxWait	Maximum wait time in minutes for background requests (1-60).	no	30
backgroundPollInterval	Initial polling interval in seconds for background requests (5-60).	no	10
skipMergeCommits	Skip merge commits during review. Merge commit diffs may not align with GitHub's PR diff, causing comment placement errors. Set to 'false' to review merge commits (may have partial success).	no	true

Outputs

no outputs