The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external content which presents an attack surface for indirect prompt injection. This is inherent to the tool's purpose of managing content.
Ingestion points: Reads content from posts.csv, ./post-content.html, and external URLs via wp media import.
Boundary markers: No specific delimiters or instructions to ignore embedded commands are included in the prompt templates.
Capability inventory: Extensive capabilities including WP-CLI command execution, file system interaction (cat, scp), and network requests via curl.
Sanitization: The skill instructions do not specify sanitization or validation logic for the content being processed.
[COMMAND_EXECUTION]: The skill executes various shell commands including wp, curl, scp, and xargs. These are necessary for the skill's stated purpose of WordPress site management.
[DATA_EXFILTRATION]: Network operations are performed via curl and wp-cli to interact with remote WordPress sites. This is consistent with the primary use case of the skill and uses placeholder credentials for authentication examples.

wordpress-content