The Agent Skills Directory

[COMMAND_EXECUTION]: The skill employs dynamic execution patterns in its validation scripts. Specifically, scripts/validate.sh uses the eval command and scripts/validate.ps1 uses Invoke-Expression to run project-specific validation tasks.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it executes commands derived from external project files without sanitization.
Ingestion points: Project configuration files including package.json, pyproject.toml, requirements.txt, go.mod, and Cargo.toml.
Boundary markers: None identified; instructions do not contain delimiters to ignore embedded commands in the target project.
Capability inventory: The skill can execute arbitrary shell commands via eval and Invoke-Expression, and uses npm, npx, pytest, go, cargo, and the GitHub CLI (gh).
Sanitization: No validation or escaping is performed on the command strings extracted from project configuration files before execution.
[COMMAND_EXECUTION]: The skill automatically generates and submits pull requests using the gh command-line tool, which involves executing external processes with potentially sensitive project data.

apex-tier2