The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on a provided shell script, wp-rest.sh, to execute WordPress and WooCommerce management tasks.
Evidence: The script performs various operations using curl, jq, and standard Unix utilities to interact with remote APIs.
[EXTERNAL_DOWNLOADS]: The check-links command in the wp-rest.sh script performs outbound HTTP requests to external URLs discovered within site content.
Evidence: The script extracts URLs from post content and uses curl to verify their status codes.
[CREDENTIALS_UNSAFE]: The skill requires the configuration of sensitive environment variables, including WP_APP_PASSWORD, WC_CONSUMER_KEY, and WC_CONSUMER_SECRET.
Evidence: Documentation in README.md and SKILL.md instructs users to provide these credentials for API authentication.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it retrieves untrusted data from the managed WordPress site that the agent then processes.
Ingestion points: Untrusted data enters the agent context through commands like list-comments, pending-comments, and get-post in wp-rest.sh.
Boundary markers: There are no explicit boundary markers or instructions to the agent to ignore commands found within retrieved site content.
Capability inventory: The skill has powerful capabilities, including delete-post, bulk-delete-old, wc-delete-product, and update-post via wp-rest.sh.
Sanitization: Content is processed via jq and sed, but raw comment and post text is passed back to the agent for evaluation.

pinch-to-post