The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the @58pic/cli global package from the official npm registry and references the author's GitHub repository (58pic-open/cli). These are documented vendor resources necessary for the skill's intended functionality.
[COMMAND_EXECUTION]: The skill operates by executing the 58pic shell command to perform searches, retrieve model lists, and initiate image generation tasks. It includes safety best practices such as recommending dry-run for potentially billable actions and using structured JSON output for agent processing.
[CREDENTIALS_UNSAFE]: The skill manages authentication via an API Key. It provides clear instructions to the agent and user to avoid hardcoding keys in repositories and suggests using the 58PIC_API_KEY environment variable or local configuration files to minimize exposure.
[PROMPT_INJECTION]: The skill processes data from the 58pic AI platform (such as search results and model descriptions), creating an indirect prompt injection surface.
Ingestion points: Data enters the agent context via the output of 58pic search, 58pic catalog, and 58pic download commands.
Boundary markers: Absent; the instructions do not specify the use of delimiters or 'ignore' instructions for the data returned from the shell commands.
Capability inventory: The agent has the ability to execute the 58pic CLI and perform general file and shell operations.
Sanitization: Absent; there is no mention of filtering or sanitizing the content retrieved from the external API before it is processed by the agent.

58pic