The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional markdown and configuration files. It does not contain any executable scripts, binary files, or commands that interact with the network or sensitive system resources.
[PROMPT_INJECTION]: A deterministic detector flagged the phrase "ignore previous instructions" in references/prompt-families.md. This is a false positive; the text is part of an educational example illustrating how models can be vulnerable to injection when input delimiters are missing, and it instructs the user to use fences to prevent such issues.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted user data ("brain dumps"). The risk is mitigated by explicit instructions in the AI Self-Check section and references/prompt-families.md to use boundary markers, XML tags, and clear labeling for all untrusted source text to prevent it from being interpreted as instructions.
[DATA_EXFILTRATION]: The skill defines local file paths for saving generated prompts (e.g., docs/local/prompts/), but it does not contain hardcoded credentials or any network operations to send data to external servers.

prompt-generator