ServiceNow Agent

Read-only CLI access to ServiceNow Table, Attachment, Aggregate, and Service Catalog APIs; includes schema inspection and history retrieval (read-only).

Before installing, verify and correct the credential metadata mismatch: the skill needs SERVICENOW_DOMAIN, SERVICENOW_USERNAME, and SERVICENOW_PASSWORD (or you must pass them on the command line). Use a least-privileged ServiceNow account scoped to read-only access and avoid saving production passwords in plaintext .env files — consider using runtime flags or a secrets store instead. Inspect cli.mjs yourself (it is included) to confirm only GET requests are made in your use cases. Note the package includes OpenAPI reference files that mention POST/DELETE endpoints; those are not used by the CLI but are an inconsistency to be mindful of. If you do proceed, run the tool in an isolated/test environment first and rotate or revoke the credentials after testing. If you need higher assurance, ask the publisher to update registry metadata to declare required env vars and to document why reference files include write endpoints even though the CLI is read-only.

Type: OpenClaw Skill Name: servicenow-agent Version: 0.1.1 The skill is designed for read-only access to ServiceNow APIs, a claim consistently reinforced in the `SKILL.md` documentation and strictly enforced by the `cli.mjs` code. While the underlying OpenAPI specifications (`references/*.yaml`) define POST, PUT, and DELETE endpoints, the `cli.mjs` script explicitly implements and allows only GET operations for all API interactions, preventing any write or modification actions. File system access is limited to reading `.env` and batch JSON files, and writing downloaded attachment content to a user-specified path, which is a legitimate function for a download utility. There is no evidence of prompt injection attempting to subvert the agent's behavior, data exfiltration to unauthorized endpoints, or malicious execution.