pscale-audit-log
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
pscalecommand-line tool to list audit logs and download authentication attempt reports. This is the primary function of the skill and targets a well-known service. - [INDIRECT_PROMPT_INJECTION]: The skill reads and processes audit log data from PlanetScale, which can contain untrusted input such as usernames or failure reasons provided by external actors during authentication attempts.
- Ingestion points: Output from
pscale audit-log listand contents of downloaded authentication export archives. - Boundary markers: The skill instructions advise confirming parameters before broad exports and warn against printing raw data to chat, but no specific prompt boundary markers are provided for the agent to distinguish data from instructions.
- Capability inventory: The skill has the capability to execute shell commands (
pscale) and write files to the local file system. - Sanitization: The instructions request output in structured formats (
json,jsonl,csv), but do not explicitly detail sanitization of values before they are presented to the agent.
Audit Metadata