Loading ToolBench
DEPRECATED — the official GEN MCP is the hosted server at https://mcp.gen.pro. Python/FastMCP port of the TypeScript server providing access to GEN's Autonomous Social Media Agent platform for content generation, idea creation, vidsheet management, and publishing.
The GEN MCP server defines 13 tools with clear naming conventions (all start with 'gen_' prefix + action verb). However, the server is DEPRECATED and officially replaced by a hosted endpoint at https://mcp.gen.pro. Input schemas are visible and properly typed with Pydantic Field annotations. Descriptions are present but vary significantly in quality, many are generic setup instructions rather than LLM-optimized guidance. Critical weakness: tool output schemas are NOT documented anywhere in the source code, and the final tool definition is truncated (source cuts off mid-parameter for gen_create_agent). Descriptions average ~150 chars, within acceptable range, but lack actionable error recovery guidance and interdependency hints. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels being assigned in the evaluation metadata. Parameter descriptions exist but are minimal (average ~40 chars) and lack format constraints, ranges, or validation rules.
Step 1 (Agent Setup): Create a new agent inside a workspace. After creation, use gen_update_agent_core to fill in identity, overview, personality, voice, and inspiration sources.
Step 1 (Agent Setup): Create a new organization/workspace. You become owner automatically. Only needed if the user doesn't already have a workspace.
Step 1 (Agent Setup): Soft-delete an agent (requires owner/manager role or being the creator).
Step 1 (Agent Setup): Permanently delete an organization and all associated data (requires owner role, irreversible).
Step 1 (Agent Setup): Get full details of a specific agent by ID. For reading full setup state (identity + overview + personality + voice + inspiration + accounts), prefer gen_get_agent_core.
Output schemas not documented. No tool response structure visible in source code. LLMs cannot plan downstream tool chains or extract necessary fields without documented return types.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels. Tools marked DESTRUCTIVE (delete_organization, delete_agent) lack confirmation patterns or dry-run options. LLMs cannot reason about irreversibility.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | D | 59 | 2026-07-28+ | v2 |
Step 1 (Agent Setup): Get the authenticated user's profile and workspace memberships. Always call first to verify your PAT and discover which workspaces you have access to.
Step 1 (Agent Setup): Get details of a specific organization by ID.
Step 1 (Agent Setup): List avatar images for an agent, with the primary avatar first.
Step 1 (Agent Setup): List agents, optionally filtered by workspace. Use the returned agent_id for ALL downstream content operations (ideas, vidsheets, generations).
Step 1 (Agent Setup): List all organizations/workspaces the authenticated user is a member of, including credits, role, and plan.
Step 1 (Agent Setup): List all workspaces the authenticated user has access to. A workspace is the billing container; every agent lives inside one.
Step 1 (Agent Setup): Update an existing agent's name, description, time zone, or voice keys. For richer setup updates (personality, inspiration, accounts, voice binding), use gen_update_agent_core.
Step 1 (Agent Setup): Update an organization's name (requires owner or manager role).
Parameter descriptions are minimal (~30-50 chars). Missing format constraints, validation rules, and interdependency hints. E.g., gen_create_agent accepts 'eleven_lab_api_key' and 'hume_ai_api_key' as bare strings with no guidance on valid format, required escaping, or which fields conflict.
Credentials (API keys) exposed as tool parameters. eleven_lab_api_key and hume_ai_api_key appear in gen_create_agent and gen_update_agent. These should use server-side secret injection; agent traces log all parameters, leaking secrets.
No error handling guidance. Tools return json_result(data) with no visible error classification, retry hints, or user-fixable error messages. An agent encountering a 403 has no guidance on what went wrong or what to try next.
Source code truncated. gen_create_agent parameter definition cuts off mid-parameter. Cannot verify complete schema for this critical tool.
Descriptions follow a template-driven format ('Step 1 (Agent Setup): ...') that is boilerplate rather than LLM-optimized. No guidance on WHEN to call which setup tool, in what order, or what alternatives exist if a step fails.
gen_list_agents and gen_list_agent_avatars accept pagination cursors but no 'limit' parameter. Cannot control result set size. Risk of context window exhaustion on large workspaces.
No permission-gate descriptions. Tools performing WRITE and DESTRUCTIVE operations (create_agent, update_organization, delete_agent, delete_organization) do not document required roles/permissions. LLMs have no visibility into whether they are authorized.