Coordinate embedding lifecycle and workspace configuration with a single entry point. Workspace Actions - get_workspace: Get current workspace folder and embedding status. - set_workspace: Set workspace folder with validation and optional embedding generation (projectPath required). - validate_wo...
AI agents use manage_embeddings to create or update resources in Ambiance MCP Server — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Ambiance MCP Server environment.
The tool writes configuration (set_workspace, validate_workspace, health_check with auto-fix), making it Write rather than Read. However, it does not delete data (not Destructive), execute arbitrary code (not Execute), or move money (not Financial). The severity is medium because misconfiguration could affect code analysis availability, but changes are reversible.
From the tool's definition Tool performs 'set_workspace' with optional embedding generation, 'health_check' with 'optional auto-fix', and manages workspace configuration. These are reversible modifications to workspace state and embedding settings.
Attacks that exploit this kind of access
Coordinate embedding lifecycle and workspace configuration with a single entry point. Workspace Actions - get_workspace: Get current workspace folder and embedding status. - set_workspace: Set workspace folder with validation and optional embedding generation (projectPath required). - validate_workspace: Validate a workspace path without setting it (projectPath required). Embedding Actions - status: Inspect current/stored model configuration, stats, and recommendations (projectPath required). - health_check: Run diagnostics with optional auto-fix for model mismatches (projectPath required). - create: Generate or regenerate embeddings using the active model settings (projectPath required). - update: Update embeddings for specific files or changed files (projectPath required, files optional). - validate: Inspect stored embeddings for compatibility issues and integrity problems (projectPath required). - check_stale: Identify files that need re-indexing by comparing disk vs database timestamps (projectPath required, autoUpdate optional). Project Management Actions - list_projects: Enumerate every project with stored embeddings. - delete_project: Remove embeddings for a specific project (requires projectIdentifier and confirmDeletion=true). - project_details: Deep dive into coverage, metadata, and compatibility for one project (requires projectIdentifier). Inputs - Set action to choose the workflow (defaults to get_workspace if no projectPath provided). - Provide projectPath for workspace-level and embedding actions. - Use projectIdentifier for project-specific lookups (delete_project, project_details). - Optional: autoGenerate (for set_workspace), autoFix, batchSize, includeStats, maxFiles, excludePatterns, allowHiddenFolders. Outputs - Returns structured results for the selected action. It is categorised as a Write tool in the Ambiance MCP Server MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Ambiance MCP Server MCP server in PolicyLayer and add a rule for manage_embeddings: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Ambiance MCP Server. Nothing to install.
manage_embeddings is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the manage_embeddings rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for manage_embeddings. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
manage_embeddings is provided by the Ambiance MCP Server MCP server (sbarron/ambiancemcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →