Companion Animation Pipeline MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| environment_infoB | Check the installed runner and Blender without changing configuration. |
| pipeline_initA | Create an isolated config for a supplied model; never overwrite an existing config. |
| pipeline_statusC | Read cached stages, failures, pending calibration and review requirements. |
| pipeline_runB | Start/resume stages as a background job. Poll job_status; exit 2 means work/review remains. |
| blender_scriptB | Run a trusted local Blender Python script in a hidden background process. The script must be inside workspace; scripts have normal user permissions, not a security sandbox. Author to derivative paths; never overwrite the source. sys.argv after -- contains the workspace path. |
| job_statusB | Read background job completion and its log tail. Does not infer visual approval. |
| record_reviewC | Record real candidate-bound visual review from an existing evidence JSON file. |
| diagnose_stageC | Record an established failure cause before retrying a stage. Does not relax tolerances. |
| view_artifactB | Display an existing PNG/JPEG inside a model workspace; maximum 10 MiB. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 9 tools
Tools are mostly distinct, but pipeline_status and job_status both report status at different scopes, and blender_script vs pipeline_run may be confused by an agent deciding between custom scripting and managed stages. Descriptions clarify the intended separation, but the overlap prevents perfect clarity.
All names use snake_case consistently, with readable prefixes such as pipeline_* and job_*. However the set mixes noun-first names (environment_info, pipeline_status) with verb-first names (record_review, diagnose_stage, view_artifact), so the pattern is not fully uniform.
Nine tools is well-scoped for an animation pipeline: setup, execution, monitoring, review, diagnosis, and artifact viewing each have a clear place. The count avoids both thinness and bloat.
The core pipeline lifecycle is covered, but pipeline_status references pending calibration while no tool records or resolves calibration. There is also no explicit stop/cancel for background jobs or cleanup/reset operation, creating potential dead ends.