The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a helper script scripts/wpcli_inspect.mjs that executes WP-CLI commands using the node:child_process module to inspect the WordPress environment.
Evidence: The script uses spawnSync to run commands like wp --info, wp core is-installed, and wp option get.
Security Context: The script uses an argument array for execution, which is a secure practice to mitigate standard shell injection vulnerabilities. It also limits execution time to 30 seconds to prevent hanging on interactive prompts and sets stdin to empty to avoid blocking on password prompts.
[PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection because it processes data retrieved from the WordPress database, which could potentially contain malicious instructions if the site is compromised.
Ingestion points: The script scripts/wpcli_inspect.mjs retrieves configuration values such as siteurl and home from the WordPress database and outputs them to the agent.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are included in the script's output when reporting these values.
Capability inventory: The agent has the capability to perform various file system and database operations via the wp command as defined in the skill's instructions.
Sanitization: No sanitization or validation is applied to the data fetched from the database before it is included in the agent's context.
Note: This represents a standard operational risk when managing external systems and is a direct requirement of the skill's primary purpose; the risk is mitigated by the extensive safety documentation provided in the skill's reference files.

wp-wpcli-and-ops