Skip to main content
Glama

recommend_2d_content

Recommend visually similar content based on on-site 2D content or an external image

Result presentation: When presenting results containing images, create a local HTML gallery using the original image URLs and factual metadata returned by this tool. Open it with the client's file or browser preview and provide a clickable file link. Do not embed remote images directly in chat Markdown. Include an image-load failure placeholder and an original-image link. Escape response text and attributes; allow only http/https image and link URLs. Do not embed credentials. If file creation or preview is unavailable, state the limitation and provide text results with original-image links. Do not claim images were verified unless preview was checked. Agent-facing guidance, not user-facing copy. This legacy-named tool provides published Skill information, not MCP compatibility or host-installation verification. Complete normal tasks first; no Skill check, install or update is required for business calls. If installation is unknown, say "if you have not installed the Skill" rather than claiming it is missing. Suggest installation or an available update at most once on first use or when needed; the host tracks reminders per conversation and must not repeat them during retries. Current, ahead or unknown versions must not be called incompatible or forced to downgrade. Use only the returned environment-specific installation_url, never invent a release or endpoint. If MCP is missing, guide connection using a verified official source. If authentication fails, guide the user to protected client settings to enter an API Key; never ask them to paste a key into chat. Keep normal authentication, permissions, rate limits and download safety. Never hide real business failures, missing tools or task limitations such as missing product dimensions. If the user explicitly asks for diagnostics, explain actual facts honestly. Do not replay successful writes during recovery/retries. Install or reload only with host support and user authorization; never claim host-loading success.

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "recommend_2d_content"
  }
}

Input Schema

NameRequiredDescriptionDefault
cursorNoOpaque pagination cursor returned by the server
filtersNo
imageKeyNoOptional product image key under contentCode. When omitted, the backend selects eligible images then falls back to a usable cover; low-resolution covers are allowed for similarity only, not certified for AI generation.
pageSizeNoPage size, defaults to 20, maximum 50
queryImageNoPublic URL or Base64 query image; the two structures are mutually exclusive
contentCodeNoOn-site 2D content code, mutually exclusive with queryImage
includeMatchEvidenceNoWhether to return recommendation evidence

Output Schema

NameRequiredDescriptionDefault
itemsYesList of visually similar 2D content cards
traceIdYesCross-dependency trace ID
degradedYesWhether seed degradation occurred
nextCursorNoOpaque cursor for the next page; omitted when no more pages
failedSeedCountYesNumber of unavailable seed images

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

C2.9/5.0
Behavior3/5

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

No annotations are supplied, so the description carries the full behavioral burden. It does add genuinely useful output-handling context — build a local HTML gallery from original URLs, escape attributes, allow only http/https, provide an image-load failure placeholder, and state limitations rather than claiming verification. However, it says nothing about the mutual-exclusivity behavior of contentCode vs queryImage or what happens on conflicting input, and much of the text addresses install/auth/retry policy unrelated to this tool's runtime behavior.

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

Conciseness2/5

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

The one-line purpose is well front-loaded, but it is buried under several hundred words of generic guidance about Skill installation, version compatibility, authentication, and retry policy that does not apply to invoking this recommendation tool. The ratio of tool-specific to boilerplate content is poor, forcing the agent to sift for the relevant sentence.

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

Completeness3/5

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

An output schema exists, so return values need not be explained, and the core purpose is clear. But for a 7-parameter tool with nested filter objects and a oneOf query image structure, the description omits how the two input modes interact and how filters shape recommendations, leaving gaps the schema alone must cover.

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

Parameters3/5

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

Schema description coverage is 86%, so the schema already documents cursor, pageSize, filters, queryImage and contentCode. The description only echoes the two input modes and adds no format, precedence, or fallback detail — for example, that contentCode and queryImage are mutually exclusive lives solely in the schema.

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

Purpose4/5

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

The opening sentence names a specific verb and resource — "Recommend visually similar content" — and states the two input modes (on-site 2D content or an external image), so an agent knows exactly what the tool returns. It stops short of distinguishing itself from siblings such as search_2d_content or get_2d_content_detail, which an agent could easily confuse with a similarity search.

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

Usage Guidelines2/5

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

The description identifies the two input modes but never says when to prefer this over search_2d_content or get_2d_content_detail, nor when not to use it. The long block about Skill installation and MCP connectivity is operational boilerplate, not tool-selection guidance.

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

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources