The Agent Skills Directory

[SAFE]: The skill provides structured instructions for a short-video hook specialist, focusing on content diagnosis and hook generation based on psychological principles.
[DATA_EXPOSURE]: The skill references local paths in a knowledge vault ('{vault}'). This is a standard architectural pattern for AI agents with context-aware memory and does not involve exfiltration to external domains.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user content (scripts/topics). While this provides an attack surface, the skill's rigid diagnostic workflow and rule-based constraints mitigate the risk of accidental obedience to embedded instructions.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns, such as the use of 'curl' or 'wget' to download and execute scripts, were found. Documentation of validation scripts in evaluation files refers to internal development processes.
[METADATA_POISONING]: All metadata fields accurately reflect the skill's purpose and functionality.

aikey-hook