Selvage

Install

Works with Claude Code, Codex, VS Code, Cursor, Windsurf, Claude Desktop. Read the repository before you give any tool access to your data.

npx agentmarket install connector/selvage-lab-selvage

About

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this server can do and what it asks for, from its tool list and install config. Checked Oct 25, 2026.

Tools

9 read · 0 write · 0 deleteRead-only

Secrets

None asked for

Install scripts

Not scanned yet

Runs as

Python package

Tools (9)

Read from the source code; the running server may differ. Descriptions are the server’s own words; hidden characters are shown as [U+…].

• get_available_models

  Get list of available AI models in Selvage. Returns: list[ModelInfo]: List of available model information with the following attributes: - name: Model name - provider: Provider (openai, anthropic, google, openrouter) - display_name: Display name - description: Model description - cost_per_1k_tokens: Cost per 1000 tokens (USD) - max_tokens: Maximum token count - supports_function_calling: Function calling support

• get_file_review_context

  Retrieve the review context for a specific file from a split context. Use this only when get_review_context returned a context_id (meaning the full context exceeded the size limit). The calling agent should call this tool in parallel for each file in the file_list returned by get_review_context. Args: context_id: Context ID returned by get_review_context file_path: File path to retrieve (from file_list) Returns: FileReviewContextResult: - success: bool - file_path: str (requested file path) - review_target: dict (the file's review context with role + cont…

• get_review_context

  Generate structured code review context for agent-delegated review. No API key required. The calling agent performs the review using its own LLM. Unlike review_* tools which call an external LLM API and return finished results, this tool returns the raw review context (system prompt and file diffs with AST-based smart context) so the agent can review the code directly in free-form text. Args: repo_path: Git repository path (default: current directory) diff_scope: Diff scope - "unstaged" (default), "staged", "branch", or "commit" target_branch: Target branch for "branch" scope (e.…

• get_review_details

  Get detailed information of a specific review. Args: log_id: Log ID of the review to retrieve Returns: ReviewDetailsResult: Review details result with the following attributes: - success: Review retrieval success status - data: Review response data (on success) - error_message: Error message (on failure)

• get_review_history

  Get recent code review history. Args: limit: Number of history items to retrieve (max 50) repo_path: Repository path filter - "CURRENT_PROJECT": Current project logs only (default) - "ALL": All project logs - Specific path: Logs for that project path (e.g., "/path/to/project") model_filter: Filter by specific model (optional) Returns: list[ReviewHistoryItem]: List of review history items with the following attributes: - log_id: Log ID - timestamp: Review timestamp (ISO 8601 format) - model: Model used for review …

• get_server_status

  Get current MCP server status. Returns: ServerStatus: Server status information

• review_changes

  Review code changes using an independent LLM API call. Requires an API key (OPENAI_API_KEY, ANTHROPIC_API_KEY, etc.). If no API key is available, use get_review_context instead to let the agent perform the review directly. Args: model: AI model to use (e.g., claude-sonnet-4, gpt-4o) repo_path: Git repository path (default: current directory) diff_scope: Diff scope - "unstaged" (default), "staged", "branch", or "commit" target_branch: Target branch for "branch" scope (e.g., "main") target_commit: Target commit hash for "commit" scope (e.g., "abc1234") Returns: ReviewRe…

• validate_api_key_for_provider

  모델의 API 키 검증 (OpenRouter-first 전략 적용) Args: model: 검증할 모델 이름 Returns: ApiKeyValidationResult: API 키 검증 결과

• validate_model_support

  모델 지원 여부 및 프로바이더 정보 검증 Args: model: 검증할 모델 이름 Returns: ModelValidationResult: 모델 지원 여부 검증 결과

Signals

Stars

36

Forks

8

Open issues

0

Contributors

2

Weekly downloads

30

Last commit

Jun 26, 2026 (4 months ago)

Last release

Feb 17, 2026 (8 months ago)

Last package publish

Feb 17, 2026 (8 months ago)

From GitHub (fetched Oct 24, 2026) and the package registry.

Details

Registry name

io.github.selvage-lab/selvage

Version

0.4.1

Auth

No auth

Packages

pypi selvage

License

Apache-2.0

Repository

github.com/selvage-lab/selvage

First seen

Oct 24, 2026

Last checked

Oct 24, 2026

Source

Official MCP Registrygithub-graphql