Skip to main content
Glama

Vixdeo — Tu Productor IA

by com.vixdeo
OverviewDiscussionsAdmin

Direct the piece

direct
Destructive

A natural-language instruction to the project's Director — the same chat turn the human director types: the planner reads the instruction, picks the ops (rewrite_scene / set_scene_fields / set_visual / swap_voice / replan_scenes / tighten_hook) and applies them to the staged draft. Suited to an editorial change that spans several fields or scenes («make scene 3 quote the source verbatim and shorten the closing»); a single exact op has its own tools. The call waits up to 60 s for the turn (50 of the 58 turns measured finished within that) and answers with the reply, the staged ops and the coherence report (outcome: completed_inline); a longer turn keeps running and the answer is its job (outcome: still_running, runId), which the project status reports by run_id. An identical call while the turn runs, or within 15 minutes after it succeeded, returns the same job, so a retry does not pay a second turn; request_id names a deliberately repeated request. One authoring turn at a time per project (authoring-in-progress). The staged draft becomes a version when it is committed. Requires write. Spends LLM budget (charged like a chat turn); a turn whose planned edits all fail stages nothing and is not charged (editOutcome: none_applied).

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
messageYesThe instruction to the Director, in any language.
qualityNoReasoned authoring pass (slower, better). Default false.
project_idYes
request_idNoOptional: a caller-chosen key for this request (up to 64 characters). The same key returns the same job for 24 h; a new key runs it again. Without it, an identical call within 15 minutes is the same request.
active_scene_indexNoOptional: the scene the instruction refers to (0-based).

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed3 schema fields changed
    • changedInput schema / properties / active_scene_index / description
      Previous value: -"Optional: the scene your instruction refers to (0-based)."New value: +"Optional: the scene the instruction refers to (0-based)."
    • changedInput schema / properties / message / description
      Previous value: -"Your instruction to the Director, in any language."New value: +"The instruction to the Director, in any language."
    • changedInput schema / properties / request_id / description
      Previous value: -"Optional: a key of your own for this request (up to 64 characters). The same key returns the same job for 24 h; a new key runs it again. Without it, an identical call within 15 minutes is the same request."New value: +"Optional: a caller-chosen key for this request (up to 64 characters). The same key returns the same job for 24 h; a new key runs it again. Without it, an identical call within 15 minutes is the same request."
  2. Changed1 schema field changed
    • addedInput schema / properties / request_id
      Added value: +{
+  "description": "Optional: a key of your own for this request (up to 64 characters). The same key returns the same job for 24 h; a new key runs it again. Without it, an identical call within 15 minutes is the same request.",
+  "type": "string"
+}
  3. First observed

TDQS

A4.8/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations only give a coarse safety profile (non-read-only, destructive, non-idempotent, closed-world); the description far exceeds that by disclosing the 60 s wait and the outcome branches (completed_inline vs still_running with runId), retry deduplication (same job within 15 min, request_id for deliberate repeats), the one-turn-at-a-time lock, and budget charging including the none_applied exemption. This is exactly the behavioral context annotations cannot convey.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with purpose and scope before dropping into timing, outcomes and budget detail; every sentence carries information. It is dense and long for a single paragraph, which costs a point, but there is little pure filler.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With no output schema, the description carries the return contract itself (reply, staged ops, coherence report, both outcome values, runId) and the state model (draft becomes a version on commit). Combined with the schema and annotations, an agent has everything needed to call it, wait for it, and interpret the result.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 80% and the schema already documents message, quality, request_id and active_scene_index. The description adds real meaning beyond the schema for request_id (names a deliberately repeated request, interacts with the 15-minute dedup window) and quality (spends LLM budget), though it does not explicitly explain active_scene_index. Slightly above the 3 baseline because the request_id semantics are genuinely enriched.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description states a specific verb+mechanism: a natural-language instruction is read by a 'planner' that picks ops (rewrite_scene / set_scene_fields / ...) and applies them to the staged draft. It explicitly distinguishes itself from siblings that handle 'a single exact op', so an agent can tell it apart from run_op/stage_op without opening schemas.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It names the situation this tool is for (an editorial change spanning several fields or scenes, with a concrete example) and points to alternatives for the opposite case ('a single exact op has its own tools'). Prerequisites (requires write, one authoring turn at a time) are also stated.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Vixdeo — Tu Productor IA toolsInspect
Try in Browser
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources