skills/jeremylongshore/claude-code-plugins-plus-skills/customerio-known-pitfalls/Gen Agent Trust Hub
customerio-known-pitfalls
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits a vulnerability surface for indirect prompt injection. The auditing logic suggests reading local TypeScript and JavaScript files to identify anti-patterns. If these files contain adversarial instructions, they could influence the agent's behavior during the audit process. \n
- Ingestion points: Local file reading (SKILL.md auditing script snippet).\n
- Boundary markers: Absent.\n
- Capability inventory: The skill requests 'Write', 'Edit', 'Bash(kubectl:)', and 'Bash(curl:)' capabilities.\n
- Sanitization: No sanitization or validation of the ingested file content is performed.\n- [COMMAND_EXECUTION]: The skill requests broad execution permissions for 'kubectl' and 'curl' via Bash. While no malicious patterns were detected in the skill's text, these tools provide high-privilege access to infrastructure and network resources that may exceed the stated purpose of auditing Customer.io integration patterns.
Audit Metadata