The Agent Skills Directory

[SAFE]: The skill does not contain any malicious patterns such as prompt injection, data exfiltration, or persistence mechanisms. Its instructions are focused on improving the quality and objectivity of the agent's reasoning.
[NO_CODE]: There are no executable scripts, binaries, or configuration files included with this skill. It consists solely of markdown instructions and metadata.
[INDIRECT_PROMPT_INJECTION]: The skill involves processing existing session data to 'strip the bait' and perform fanned-out reads. Although it ingests untrusted data from the session history, it lacks any dangerous capabilities such as network access, file system modification, or command execution, rendering the risk of exploitation negligible.

macrothink