The Agent Skills Directory

[SAFE]: No prompt injection or behavior override markers were found. The instructions strictly define a persona and analytical framework without attempting to bypass safety filters or extract system prompts.
[SAFE]: No data exposure or exfiltration risks detected. The skill does not access sensitive local file paths (such as .ssh or .env), nor does it perform network operations to external domains.
[SAFE]: No obfuscation techniques, such as Base64 encoding, zero-width characters, or homoglyphs, were found in the instructions or reference files.
[SAFE]: No external dependencies or remote code execution patterns are present. The skill does not use package managers (pip, npm) or download remote scripts via curl or wget.
[SAFE]: No persistence mechanisms or privilege escalation attempts were identified. There are no modifications to shell profiles, cron jobs, or system-level configurations.
[SAFE]: Metadata and frontmatter are legitimate and descriptive, with no signs of metadata poisoning or deceptive instructions.

berkshire