The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional markdown content and design guidelines. It contains no executable scripts, binaries, or automated tasks that pose a direct security risk.\n- [INDIRECT_PROMPT_INJECTION]: The skill explicitly addresses the security implications of processing untrusted data from third-party platforms (webhooks and API responses). It instructs the agent to mitigate risks by verifying cryptographic signatures, fetching objects by ID rather than trusting webhook bodies, and escaping user-supplied values in HTML. These instructions effectively manage the ingestion surface for malicious payloads.\n- [CREDENTIALS_UNSAFE]: The skill provides constructive guidance on secure credential management, recommending database locks for token renewals and monitoring for expiration. It does not contain any hardcoded secrets or sensitive environment configuration.\n- [COMMAND_EXECUTION]: No shell commands, dynamic context injection, or risky subprocess patterns were found. The skill remains within the bounds of development guidance.

integrations