The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides various scripts and Makefiles designed to execute system-level commands for building, testing, and deploying applications.
Evidence: scripts/dev-setup.sh executes docker-compose and npm commands. scripts/deploy.sh uses ssh to execute commands on remote production and staging environments.
[INDIRECT_PROMPT_INJECTION]: The provided automation scripts represent a surface for indirect injection if the data sources they ingest (such as environment variables or command-line arguments) are sourced from untrusted environments.
Ingestion points: scripts/dev-setup.sh reads from .env files and user prompts; scripts/deploy.sh accepts environment labels as arguments.
Boundary markers: No explicit sanitization or delimiters are used in the provided templates.
Capability inventory: The scripts have capabilities to perform file system operations, network access via ssh, and package management via npm across SKILL.md.
Sanitization: The templates do not include input validation logic, relying on the user to provide valid configuration.
[EXTERNAL_DOWNLOADS]: The GitHub Actions workflow integrates with well-known third-party services for development tasks.
Evidence: .github/workflows/ci.yml uses codecov/codecov-action@v3 to upload code coverage reports.

workflow-automation