context-engine-ai/Context Engine Reviewer

AI-powered PR reviews for Context Engine workflows.

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Stale
License
None
namedescriptionrequireddefault
style_guide_rulesCustom style guide rules that will be enforced during review by generating critical commentsno""
github_api_urlGitHub API URL for GitHub Enterprise Server (e.g., https://github.example.com/api/v3)nohttps://api.github.com
github_server_urlGitHub Server URL for GitHub Enterprise Server (e.g., https://github.example.com)nohttps://github.com
custom_modeEnhanced senior engineer review mode: on | off | autonoauto
llm_modelOverride LLM model (e.g., gpt-4o-mini)no""
llm_providerOverride LLM provider (e.g., ai-sdk)no""
llm_reasoning_effortReasoning effort for models that support it (e.g., low | medium | high). Claude Fable is always pinned to medium.no""
zai_base_urlZ.AI OpenAI-compatible base URL for glm-* modelsnohttps://api.z.ai/api/coding/paas/v4/
kimi_base_urlKimi (Moonshot AI) OpenAI-compatible base URL for kimi-* and moonshot-* modelsnohttps://api.moonshot.ai/v1
deepseek_base_urlDeepSeek OpenAI-compatible base URL for deepseek-* modelsnohttps://api.deepseek.com/v1
context_engine_api_keyOptional Context Engine API key. When set, reviewer LLMs can call Context Engine MCP tools for repository context.no""
context_engine_mcp_urlOptional Context Engine MCP indexer URL.nohttps://dev.context-engine.ai/indexer/mcp
context_engine_collectionOptional Context Engine collection name to scope reviewer MCP searches.no""
context_engine_toolsOptional comma-separated Context Engine MCP tools exposed to the reviewer LLM.norepo_search,batch_search,symbol_graph,batch_symbol_graph,graph_query,batch_graph_query,search_tests_for,search_config_for,search_commits_for,context_answer
context_engine_max_toolsMaximum number of Context Engine MCP tools exposed to the reviewer LLM.no10
context_engine_max_stepsMaximum LLM steps (tool roundtrips plus final output) per review batch when Context Engine tools are enabled.no8
review_scopesComma-separated review scopes to run: security,performance,best-practicesno""
max_commentsMaximum inline comments to post per runno40
max_codeblock_linesMaximum lines retained inside fenced code blocks in commentsno60
max_review_charsApproximate maximum diff characters per LLM review batchno725000
allow_title_updateAllow the action to update PR title when @context-engine-reviewer is present (default: false)nofalse

no outputs