The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from GitHub issues which could contain malicious instructions. While the workflow includes human gates, the data is processed by the agent without explicit boundary markers or instruction-override protection.\n
Ingestion points: SKILL.md (Step 1.1) retrieves issue details using gh issue view.\n
Boundary markers: The skill does not implement explicit delimiters or warnings to ignore instructions within the retrieved issue text.\n
Capability inventory: The skill executes powerful tools including git, terraform, gcloud, gh, and uv for local and cloud operations.\n
Sanitization: No sanitization is performed on ingested issue content for potential prompt injections, though PII sanitization is mandated for PR outputs.\n- [COMMAND_EXECUTION]: The skill executes various command-line utilities (git, gh, terraform, gcloud, uv) to manage the development lifecycle, verify infrastructure, and interact with repository hosting services.\n- [EXTERNAL_DOWNLOADS]: Fetches provider documentation from HashiCorp's official GitHub repository using curl to verify resource configurations.

contributing