ClamAV MCP server for scanning files using ClamAV antivirus engine
Single tool with basic schema and description. Tool name 'scan_file' is appropriately verb-noun formatted. Description is present (67 chars) but lacks critical context: no mention of prerequisites (ClamAV must be installed), no guidance on when to use it, no explanation of return structure. Input schema has proper types and descriptions for both parameters. However, output schema is undocumented, LLM cannot predict response structure. No error handling guidance; generic exception returns provide no recovery path. No parameter constraints (e.g., max base64 size, filename format). Missing security considerations around file handling and temporary file cleanup.
Scan a base64-encoded file using ClamAV.
Output schema undocumented. LLM cannot predict response structure (success flag, result field, error field). Breaks downstream tool chaining and forces LLM to guess field names.
Description lacks prerequisites and context. Does not state that ClamAV must be installed/running, what 'success' means, or when to call this tool. LLM cannot determine if tool is available.
No error handling guidance. Generic exception returns ('success': false, 'error': str(e)) provide no recovery path. LLM cannot distinguish retryable errors (timeout, temp file creation) from fatal ones (ClamAV not found).
No input constraints. base64_data parameter accepts unbounded strings; no max size specified. LLM could pass multi-GB base64 strings, causing memory exhaustion or timeout.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 50 | <=2025-11-25 | v2 |
No timeout handling. subprocess.run() has no timeout parameter. Hung ClamAV process blocks the entire agent indefinitely.