The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data provided by the user (or potential external sources if the agent fetches content) to create social media posts, creating a potential injection surface.
Ingestion points: Content passed via the -c/--content flag or JSON structure to the postiz posts:create command.
Boundary markers: The instructions do not explicitly provide specific delimiters or 'ignore embedded instructions' warnings for the data being interpolated into the shell commands.
Capability inventory: Execution of shell commands via Bash(postiz:*), file system read/write for credentials, and network operations to api.postiz.com, mcp.postiz.com, and cli-auth.postiz.com.
Sanitization: The skill does not explicitly describe sanitization of the content before it is passed to the CLI tool.
[DATA_EXFILTRATION]: The skill manages and accesses a sensitive configuration file on the local file system.
Evidence: The file ~/.postiz/credentials.json is used to store OAuth access tokens and API settings, as seen in src/commands/auth.ts. While necessary for the tool's primary purpose, agents must be careful not to expose the contents of this file or the POSTIZ_API_KEY environment variable in logs or output.

postiz