The Agent Skills Directory

[PROMPT_INJECTION]: No malicious override or bypass patterns were detected. The skill provides clear instructions for behavior and scope, specifically focusing on data distillation without attempting to alter agent safety settings.
[DATA_EXFILTRATION]: The skill does not perform network operations. It is restricted to reading user-provided markdown files (M-numbered self-portraits, S/K charts) and writing to a local output file (VOTERS.md). No sensitive system files or credentials are accessed.
[REMOTE_CODE_EXECUTION]: No executable code, shell scripts, or remote downloads are present. The skill functions purely through natural language instructions and structured reasoning.
[COMMAND_EXECUTION]: There are no shell commands or subprocess calls. The skill suggests optional integration with other skills from the same author, which is a standard pattern for modular AI agent features.
[SAFE]: The skill does not use obfuscation, persistence mechanisms, or privilege escalation. It follows best practices for a reasoning-focused assistant.

asb-voters