Skip to main content
Glama

The Clearing — The Lantern & Consensual Agent Lab

audit_multimodal_prompt_hygiene

Audit an image/audio caption, OCR snippet, or SVG text layer for indirect prompt injection, QR/URI exfil lures, and Zero-PII compliance.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "text_content": "<text_content>"
    },
    "name": "audit_multimodal_prompt_hygiene"
  }
}

Input Schema

NameRequiredDescriptionDefault
text_contentYesExtracted OCR, caption, or VTT text snippet.
source_channelNoOptional modality source: ocr, vllm, qr, audio_vtt, or cua.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

B3.2/5.0
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 disclose the audit targets (prompt injection, QR/URI exfil lures, Zero-PII), which is meaningful behavior beyond a bare name, but it omits whether the operation is read-only and side-effect free, whether it blocks or only reports, and what a finding looks like. That leaves real gaps for a security tool with zero annotation coverage.

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?

A single front-loaded sentence with no filler; the audited inputs and the threat classes are stated up front. It is dense with specialized terms (Zero-PII, exfil lures) but every clause earns its place.

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?

For a two-parameter, no-output-schema, no-annotation tool, the description is adequate on the 'what is checked' axis but incomplete on 'what comes back' and 'how the optional channel parameter affects results'. An agent could invoke it correctly for basic cases but lacks enough to interpret or act on findings.

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 coverage is 100% and both parameters are already documented, so the baseline is 3. The description adds some mapping between modalities (image/audio caption, OCR, SVG) and the source_channel values, but it does not clarify that source_channel is optional, how channel selection changes detection, or what the unmentioned 'cua' channel means.

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 description names a specific verb ("Audit") plus concrete resources (image/audio caption, OCR snippet, SVG text layer) and three explicit targets: indirect prompt injection, QR/URI exfil lures, and Zero-PII compliance. It does not differentiate from close siblings such as audit_multimodal_schema_fidelity or inspect_svg_multimodal_surface, so an agent cannot tell from the text alone which sibling handles SVG inspection versus hygiene auditing.

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?

There is no when-to-use or when-not-to-use guidance and no named alternative. The only implicit cue is the content types listed, which hints the tool applies to untrusted multimodal text, but nothing states prerequisites (e.g., that text must first be extracted) or which sibling to pick for schema or branding checks.

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