The Agent Skills Directory

[SAFE]: The skill consists of a long-form instructional prompt for the AI to act as a career counselor. It uses standard role-playing techniques and structured dialogue rules (Socratic questioning, multi-round interaction) to achieve its goal.
[SAFE]: The SKILL.md frontmatter contains disable-model-invocation: true, which restricts the agent's environment, reducing the attack surface.
[PROMPT_INJECTION]: The skill contains instructions like 'Ignore your own process' and 'Disregard changes,' but these are within the context of maintaining the integrity of the specific counseling persona and are not malicious attempts to bypass safety filters or extract system prompts.
[DATA_EXPOSURE]: No sensitive file paths, environment variables, or hardcoded credentials were found.
[EXTERNAL_DOWNLOADS]: No external network requests or remote scripts are executed.
[COMMAND_EXECUTION]: There are no shell commands or subprocess calls present in the instructions or configuration files.

hidden-talent