The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from GitLab issue descriptions and external comments, which are potentially controlled by untrusted parties. This creates an indirect prompt injection surface. However, this is inherent to the skill's primary function of requirements enrichment, and the risk is mitigated by strictly defined output templates and constrained logic.
[SAFE]: Implements secure communication by default, using the internal: true flag when posting GitLab issue comments to ensure internal-only visibility.
[SAFE]: Employs strict identity verification for mentions, requiring the agent to use only the author.username returned from the official GitLab MCP tool, which prevents impersonation or metadata-based spoofing.
[SAFE]: No evidence of unauthorized command execution, remote code downloads, or network exfiltration was found in the analysis of the skill or its associated reference files.

ns-requirements-enricher