The Agent Skills Directory

[PROMPT_INJECTION]: No attempts to override safety filters or bypass platform guidelines were detected. The skill includes clear negative constraints forbidding the model from creating facts or judging user market value.
[DATA_EXFILTRATION]: There are no network operations, API calls, or hardcoded secrets. The skill explicitly instructs against collecting unnecessary personal information (PII) and handles confidential data through abstraction.
[COMMAND_EXECUTION]: No dangerous shell commands or privilege escalation attempts were found. The included Python scripts use standard library modules for JSON processing and validation without executing arbitrary system commands.
[INDIRECT_PROMPT_INJECTION]: While the skill processes work history data, it uses a deterministic script (check_inventory_coverage.py) with strict type validation defined in _common.py to analyze the input, which prevents data-driven instruction injection.
[DYNAMIC_EXECUTION]: No dynamic code generation or remote script execution (e.g., curl|bash) is present. All logic is contained within local, static scripts.
[OBFUSCATION]: The content is clear and readable. No Base64, zero-width characters, or homoglyph techniques were identified.

career-inventory