The Agent Skills Directory

[SAFE]: The skill implements engineering best practices for Python development, focusing on cross-platform compatibility through pathlib and explicit character encodings.
[DATA_EXPOSURE]: The instructions explicitly direct the agent to perform security reviews for sensitive data, including hardcoded secrets and path traversal vulnerabilities, which improves the overall security posture of the generated code.
[REMOTE_CODE_EXECUTION]: The resource links in references/authoritative-sources.md point exclusively to well-known and trusted official documentation for Python packaging and GitHub Actions, representing standard development resources.
[COMMAND_EXECUTION]: The skill references standard operating system utilities for task scheduling (cron and schtasks) and environment management, which are appropriate for its stated DevOps purpose.
[INDIRECT_PROMPT_INJECTION]: As the skill is designed to ingest and review external geospatial code (specified in SKILL.md), it possesses an inherent surface for indirect prompt injection; however, this is mitigated by instructions to prioritize security reviews and the lack of high-risk capabilities like unverified remote code execution. (Ingestion points: external geospatial code via SKILL.md. Boundary markers: absent. Capability inventory: file system writing and subprocess execution for linting and testing via SKILL.md. Sanitization: absent.)

swe-devops-standards