The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses npm and the vendor-specific opendeploy CLI to perform version checks and API operations.
[DATA_EXFILTRATION]: The skill leverages local credentials stored in ~/.opendeploy/auth.json. Usage is restricted to the official vendor domain dashboard.opendeploy.dev, which is consistent with the skill's primary purpose and the author's infrastructure.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing external API responses.
Ingestion points: API response data retrieved via opendeploy api get (referenced in SKILL.md).
Boundary markers: Present; the skill explicitly instructs the agent to "Validate the response shape before acting on it" and to stop if an unrecognized structure is found.
Capability inventory: The skill allows subprocess calls via the opendeploy api tool for post, put, and patch operations.
Sanitization: No explicit sanitization or escaping is implemented in the shell-based instructions, but boundary markers are present to guide the agent.

opendeploy-api