airtable-cli
Warn
Audited by Gen Agent Trust Hub on Jul 25, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill's metadata contains a discrepancy where the frontmatter identifies the author as 'airtable' despite the skill being published by 'onmax'. This misleading identity can influence an agent's trust assessment.- [CREDENTIALS_UNSAFE]: The skill references sensitive file paths for credential storage, specifically
~/.airtable/cli.json, and instructs users on managing 'AIRTABLE_TOKEN' environment variables.- [EXTERNAL_DOWNLOADS]: Fetches and installs the@airtable/mcp-clipackage from the npm registry.- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection attacks because it handles untrusted external data. - Ingestion points: Retrieves record data and table schemas from Airtable using
list-records-for-tableandsearch-records. - Boundary markers: The instructions lack explicit delimiters or 'ignore embedded instructions' warnings for processed data.
- Capability inventory: The skill has the ability to write to Airtable via
update-records-for-tableand execute various CLI tools. - Sanitization: No evidence of input validation or content sanitization is present in the instructions.- [COMMAND_EXECUTION]: Executes shell commands using the
airtable-mcputility, including operations that take structured JSON input from stdin using the--input -flag.
Audit Metadata