The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Recommends the global installation of the apifox-cli package from a well-known mirror (registry.npmmirror.com) if the binary is not found on the system.
[COMMAND_EXECUTION]: Executes shell commands using the apifox binary to perform project operations, including resource creation, deletion, and environment management.
[CREDENTIALS_UNSAFE]: Manages API access tokens through the apifox login command and accesses sensitive configuration data stored in ~/.apifox/config.toml. The skill includes instructions to the agent to avoid leaking these credentials into logs or chat summaries.
[PROMPT_INJECTION]: The skill instructs the agent to prioritize following instructions found in the agentHints.nextSteps field of the CLI's JSON output. This represents an indirect prompt injection surface where the tool's output can dictate the agent's logic.
Ingestion points: CLI JSON output parsed by the agent in SKILL.md.
Boundary markers: Absent; the instructions explicitly tell the agent to let these hints "drive" the next steps.
Capability inventory: Full CRUD access to Apifox resources, CI/CD configuration management, and file system writes for configuration.
Sanitization: No sanitization or verification of the agentHints content is specified.

apifox-cli