The Agent Skills Directory

[SAFE]: The skill provides legitimate DevOps functionality for generating CI/CD configuration files. Analysis of the instructions and examples shows adherence to security best practices, such as recommending the use of CI/CD variables for secret management rather than hardcoding credentials.
[COMMAND_EXECUTION]: The skill employs basic discovery commands (ls, cat, grep) to detect project metadata (e.g., package.json, Makefile). These commands are used for environment analysis only and do not involve unsafe data interpolation or privilege escalation.
[INDIRECT_PROMPT_INJECTION]: The skill has a defined attack surface as it ingests local project files to determine the appropriate CI/CD template. However, the risk is minimal as the generated output consists of static YAML configuration intended for user implementation rather than direct execution by the agent.
Ingestion points: Reads from project manifests like package.json, pyproject.toml, and Makefile via cat and grep in SKILL.md.
Boundary markers: None specified for the read operations.
Capability inventory: Generates .gitlab-ci.yml and .circleci/config.yml configuration strings.
Sanitization: Not applicable; the skill produces configuration templates for manual setup.

cicd-pipeline