airtable-filters
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for the agent to transform natural language user requests into structured Airtable filter parameters, which involves processing untrusted user input for use with external MCP tools. 1. Ingestion points: User requests to find, filter, or search Airtable records as described in SKILL.md. 2. Boundary markers: The skill does not define specific delimiters or escaping mechanisms for user-provided data within the generated JSON filter objects. 3. Capability inventory: The instructions target MCP tools that list or display records from Airtable tables and interface pages via the Airtable API. 4. Sanitization: The documentation promotes structural integrity by requiring schema-derived IDs and specific date/range objects, but it does not include instructions for sanitizing the logical content of user queries.
- [NO_CODE]: The skill contains no executable code, scripts, or package dependencies, consisting entirely of instructional markdown and JSON examples.
Audit Metadata