The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface in the resources/implementation-playbook.md file. The WorkflowAnalyzer class contains methods that read the contents of untrusted repository files, including README documentation and existing CI/CD configurations, to inventory development processes. This content is processed without sanitization or boundary markers, which could allow instructions embedded in these files to manipulate the agent's behavior during the automation design phase.
Ingestion points: resources/implementation-playbook.md reads .github/workflows/*.yml, .gitlab-ci.yml, Jenkinsfile, and README.* from the project directory.
Boundary markers: Absent.
Capability inventory: The skill facilitates the design of workflows, configuration of secrets, and execution of shell scripts for environment setup.
Sanitization: Absent.
[EXTERNAL_DOWNLOADS]: The skill's implementation playbook and automation templates reference numerous external dependencies and tools. The setup scripts use package managers like npm and pip to install development utilities, while the GitHub Actions templates utilize various actions from well-known technology providers such as AWS, Snyk, and Aqua Security for testing and deployment operations.
[COMMAND_EXECUTION]: The provided scripts and templates include multiple commands for local and remote execution. This includes development environment setup via npm ci and docker-compose in scripts/setup-dev-environment.sh, and cloud infrastructure operations like aws ecs and terraform apply in the pipeline examples.

cicd-automation-workflow-automate