The Agent Skills Directory

[SAFE]: The skill implements a robust workflow that includes a 'Draft & Confirm' stage (Step 6), ensuring a human-in-the-loop verification before the agent executes the create_issue tool in Linear.
[SAFE]: The skill uses Linear and Sentry MCP tools for their intended purposes (fetching issue details and creating reports). No unauthorized or suspicious tool usage was detected.
[SAFE]: The instructions mandate the use of sub-agents (Task) to search the codebase and fetch Sentry data. This design pattern isolates potentially untrusted or verbose data (like stack traces) and filters it before it enters the main agent context, mitigating risks associated with context pollution or indirect prompt injection.
[SAFE]: No indicators of malicious behavior, such as prompt injection, data exfiltration, obfuscation, or hardcoded credentials, were found across the skill's instructions, templates, and example files.

ticket-creator