The Agent Skills Directory

[SAFE]: The skill follows a standard workflow for profile optimization without utilizing any dangerous system calls, network operations, or unauthorized file access.
[PROMPT_INJECTION]: The skill uses the 'CRITICAL:' marker and 'Auto-start' instructions to direct the agent to skip introductory explanations and begin the workflow immediately. While these instructions override standard agent conversational patterns, they are for user experience optimization rather than bypassing security protocols or ethical guidelines.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from an optional local file ('about-me.md') and user-provided profile data (text, PDF, or screenshots). This represents an attack surface for indirect prompt injection. Ingestion points: Reading of 'about-me.md' and user inputs/uploads. Boundary markers: The skill does not employ explicit delimiters or 'ignore' instructions to isolate user-provided data from its own logic. Capability inventory: The skill is limited to generating text and interacting with the user via 'AskUserQuestion'. It does not have access to the network, shell commands, or sensitive system files. Sanitization: No validation or sanitization of the input data is performed.

profile-optimizer