The Agent Skills Directory

[NO_CODE]: This skill consists entirely of markdown documentation (SKILL.md and references/experiments.md). It contains no scripts, binaries, or automated tasks that could execute code on the host system.
[SAFE]: No security threats were identified across any analyzed categories.
No hardcoded credentials or sensitive data exposure detected.
No network operations or external downloads found.
No prompt injection markers or obfuscation techniques present.
[DATA_EXPOSURE]: The skill references a local file context (.claude/product-marketing-context.md) for gathering product information. This is a standard practice for context-aware agents and does not involve exfiltration or unauthorized access.
[PROMPT_INJECTION]: The instructions are clear, descriptive, and follow a constructive guidance format without attempts to bypass system filters or override core agent safety protocols.

paywall-upgrade-cro