The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill documentation identifies that the CLI stores sensitive authentication tokens, API URLs, and dashboard URLs in ~/.openship/config.json (references/cli/access.mdx). It also references the storage of environment secrets and database passwords in ~/.openship/compose/.env (references/guides/updating.mdx). Accessing these file paths can lead to the exposure of credentials if the agent's environment is compromised.
[EXTERNAL_DOWNLOADS]: The skill instructs the agent to install the openship package via npm and references source code and release bundles hosted at github.com/oblien/openship (references/upstream.json, references/cli/index.mdx). Neither the npm package nor the GitHub repository/user is verified or included in the trusted organization list, which presents a potential supply chain risk for users installing the tool.
[COMMAND_EXECUTION]: The skill relies extensively on executing shell commands through the openship CLI to manage servers, deployments, and backups. This provides a broad capability set, including service lifecycle management and system configuration, which could be abused if the agent is directed toward malicious actions.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a significant attack surface for indirect prompt injection by ingesting untrusted external data.
Ingestion points: The skill is designed to read and follow runtime logs and server request logs (references/guides/logs-monitoring.mdx) and inspect repository source code (references/cli/deploy.mdx).
Boundary markers: There are no specific boundary markers or instructions for the agent to ignore embedded instructions within the logs or repository data it processes.
Capability inventory: The skill provides the agent with the ability to perform high-impact operations via the CLI, such as renaming projects, deleting services, and modifying environment variables (references/cli/projects.mdx).
Sanitization: The skill does not describe any sanitization or validation protocols for the external data ingested from logs or repositories.

openship-deploy