The Agent Skills Directory

[DYNAMIC_EXECUTION]: The vault.py component dynamically constructs an Ansible playbook in a temporary location to verify vault credentials. This runtime generation and execution of code is a significant capability that warrants careful review.
[COMMAND_EXECUTION]: The skill relies heavily on executing external binaries such as ansible, docker, and git via subprocess.run. These operations are fundamental to the skill's purpose but provide a broad interface for system interaction.
[PRIVILEGE_ESCALATION]: Deployment tasks are performed with administrative privileges (become: true), and the skill configures the operational user to be a member of the docker group, which grants extensive control over the host environment.
[INDIRECT_PROMPT_INJECTION]: The skill operates by reading and interpreting user-supplied project files, creating an attack surface where malicious configuration data could influence the agent's behavior.
Ingestion points: Reads Dockerfile, compose.yaml, SEMVER, and Ansible inventory files from the project root.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to distinguish between its own logic and potentially malicious instructions embedded in the project files.
Capability inventory: The skill can execute remote commands, deploy network stacks, modify local files, and synchronize SSH keys.
Sanitization: Validation is primarily focused on structural correctness rather than the semantic safety of the deployment logic.
[DATA_EXPOSURE]: The skill automatically identifies and synchronizes SSH public keys from the controller machine. While private keys are explicitly excluded, the automated handling of authentication credentials represents a sensitive data access pattern.

specsfy-specialist-deploy