Native Airtable

Read Airtable bases, tables, and records directly via the Airtable API. Use when you need spreadsheet/database data from Airtable. Calls api.airtable.com dir...

This skill appears to do exactly what it claims: call api.airtable.com using a Personal Access Token (AIRTABLE_PAT) and print results. Before installing, consider: - Use a least-privilege PAT (scopes data.records:read and schema.bases:read and limit the bases it can access). If possible create a token that only covers the bases the agent must read. - The script prints record data to stdout; anything the agent does with that output (logging, forwarding) is your responsibility. Don’t provide a PAT that exposes unrelated sensitive bases. - The code has minor correctness issues (search formula formatting and fields handling) and does not sanitize user-supplied formula/query strings — avoid running untrusted queries if you’re worried about unexpected Airtable formula behavior. - If the PAT is ever exposed or you stop using the skill, revoke the token in your Airtable account. Overall this skill is internally consistent and low-risk, but follow standard secret-management practices and review output handling in your agent environment.

Type: OpenClaw Skill Name: native-airtable Version: 0.1.0 The skill is designed to interact with the Airtable API, and its `SKILL.md` instructions are benign. However, the `scripts/airtable.py` file contains a vulnerability in the `cmd_search_records` function. User-provided `query` and `field` arguments are directly embedded into an Airtable formula string without proper sanitization, which could lead to an Airtable formula injection. This vulnerability could allow an attacker to manipulate the search query on the Airtable server, potentially leading to unintended data access or API errors, though it does not pose a direct risk of RCE or data exfiltration from the local system.