The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The troubleshooting section recommends mounting the host Docker socket (/var/run/docker.sock) into a containerized build agent to resolve connection issues. This configuration grants the containerized process full control over the host's Docker daemon, enabling a container escape to the host system and potential full host compromise. Additionally, multiple code examples specify running containers as the root user using the '-u root' argument.\n- [COMMAND_EXECUTION]: The skill relies extensively on shell execution ('sh') within Jenkins pipelines, which is the primary mechanism for build automation but provides a large attack surface if inputs are not validated.\n- [INDIRECT_PROMPT_INJECTION]: The skill assists in the creation of CI/CD pipelines which naturally ingest untrusted data from source code, pull requests, and branch names.\n
Ingestion points: Jenkinsfile pipeline definitions and project source code processed by 'sh' steps in the provided Groovy examples.\n
Boundary markers: No specific delimiters or safety instructions (e.g., 'ignore embedded instructions') are recommended for handling untrusted branch content.\n
Capability inventory: Extensive shell execution capabilities ('sh'), credential injection ('withCredentials'), and container orchestration ('kubernetes', 'docker').\n
Sanitization: While secret injection follows best practices using 'withCredentials', the skill does not define specific sanitization or escaping for external content interpolated into shell commands.

jenkins