The Agent Skills Directory

[SAFE]: The skill implements a robust security model by requiring the agent to hand over control to the user before any credentials are entered, confirmed, or submitted. This ensures that the agent never performs sensitive account modifications autonomously.
[DATA_EXPOSURE]: The instructions include strict prohibitions against requesting or handling sensitive data such as passwords, multi-factor authentication codes, or password manager vault exports. It mandates that these secrets remain within protected official flows and are never exposed in logs, screenshots, or chat history.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an ingestion surface for potentially untrusted data, as it is instructed to read browser security reports and account provider notices. However, the risk is mitigated by explicit instructions to never extract or compare secret values and to rely on manual user verification for all confirmed findings.

secure-me