The Agent Skills Directory

[SAFE]: The skill consists entirely of Markdown and JSON files containing documentation, templates, and frameworks. There are no executable scripts, shell commands, or network-enabled tools.- [PROMPT_INJECTION]: Phrases flagged as instruction overrides (e.g., 'ignore previous instructions') are explicitly documented as examples of common failure modes or as part of a 'System/User Separation Pattern' designed to prevent real-world prompt injection.- [INDIRECT_PROMPT_INJECTION]: The skill provides specific guidance on handling untrusted content, recommending the use of XML tags and explicit delimiters to isolate user-provided data from agent instructions.- [NO_CODE]: As the skill contains no executable components, it possesses no capability to perform data exfiltration, privilege escalation, or persistence.

prompt-improver