The Agent Skills Directory

[PROMPT_INJECTION]: No attempts to override agent behavior or bypass safety guidelines were detected. The instructions are purely technical and focus on Convex development best practices.
[DATA_EXPOSURE]: No hardcoded credentials, sensitive file access, or suspicious network operations were found. The use of 'fetch' in code snippets is purely illustrative and targets a well-known service (Stripe) or placeholder domains.
[REMOTE_CODE_EXECUTION]: The skill does not perform any remote script execution or download untrusted code. It uses standard Convex framework imports.
[COMMAND_EXECUTION]: No dangerous shell commands or privilege escalation attempts were identified. The allowed-tools list is appropriate for the skill's stated purpose of backend development.
[OBFUSCATION]: No hidden content, encoded strings (Base64), or zero-width characters were found in the skill content.
[INDIRECT_PROMPT_INJECTION]: The skill provides templates for data ingestion that include robust validation schemas using Convex's validation library (e.g., 'v.string()', 'v.id()'), which follows security best practices for handling external data.

convex-fundamentals