The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external data, such as GitHub issue and pull request templates, and remote source code. This creates a theoretical surface for indirect prompt injection if external content contains instructions intended to manipulate the agent's behavior. However, the skill processes this data for structural validation (e.g., matching markdown headers and checkboxes) and informational display, which minimizes security risk.
Ingestion points: Project-specific issue and PR templates (.github/ISSUE_TEMPLATE/, .github/PULL_REQUEST_TEMPLATE.md) and remote files fetched via the GitHub API (gh api).
Boundary markers: The skill recommends using explicit delimiters and disclosure statements when filing issues.
Capability inventory: File system access for reading/writing local scripts and templates; network access via git and gh CLI for interacting with GitHub repositories.
Sanitization: The provided compliance script (scripts/check-pr-template-compliance.py) uses regular expressions to validate document structure rather than executing content.
[COMMAND_EXECUTION]: The instructions and reference materials provide standard shell command examples for utilizing git and the GitHub CLI (gh). These commands are standard for the skill's purpose of managing open-source contributions and do not include high-risk patterns like privilege escalation or arbitrary command injection.
[DYNAMIC_EXECUTION]: The skill includes a standalone Python script (scripts/check-pr-template-compliance.py) that performs local file validation. The script uses standard libraries (re, sys) for text processing and does not employ dangerous functions such as eval() or exec().

opensource-contributions