The Agent Skills Directory

[PROMPT_INJECTION] (SAFE): The skill contains 'Tell AI' blocks, which are intended as user-provided templates for a language model. These do not contain override or bypass markers designed to subvert the agent's safety protocols.
[DATA_EXFILTRATION] (SAFE): No sensitive file access or network exfiltration patterns detected. The credit card numbers provided (e.g., 4242424242424242) are standard, publicly known Stripe test credentials used for development purposes.
[REMOTE_CODE_EXECUTION] (SAFE): There are no scripts, binaries, or instructions to download and execute remote code.
[OBFUSCATION] (SAFE): No hidden, encoded, or deceptive text was found. The content is plain-text markdown.
[CREDENTIALS_UNSAFE] (SAFE): No real API keys or secrets are hardcoded. The skill correctly advises users to use environment variables for actual Stripe keys.
[INDIRECT_PROMPT_INJECTION] (SAFE): The skill identifies the attack surface of webhooks but provides explicit remediation guidance, such as using stripe.webhooks.constructEvent() for signature verification to prevent spoofing.
[PRIVILEGE_ESCALATION] (SAFE): No commands related to privilege escalation (e.g., sudo, chmod) are present.

payments