The Agent Skills Directory

[SAFE]: The skill is authored by a trusted vendor (Sentry) and exclusively uses official Sentry tools and packages for instrumentation. All installation and configuration steps align with standard developer workflows.\n- [SAFE]: Security instructions are explicitly provided in SKILL.md, directing the agent to treat data from the Model Context Protocol (MCP) as untrusted input and to avoid executing any instructions found within event payloads or comments. This effectively mitigates the risk of indirect prompt injection from monitored application data.\n- [SAFE]: The skill promotes secure handling of sensitive information, such as Data Source Names (DSNs) and authentication tokens, by recommending the use of environment variables and secrets management systems rather than hardcoding credentials.\n- [SAFE]: Setup verification steps in references/setup-verification.md include recommendations for cleaning up temporary test triggers once the integration is confirmed, preventing the introduction of unnecessary code paths in the user's application.

sentry-get-started