fermionoid
@fermionoid
Public Skills
senseguard
by fermionoid
Semantic security scanner for OpenClaw skills. Detects prompt injection, data exfiltration, and hidden instructions that traditional code scanners miss. Use when user asks to scan skills, check skill safety, or run a security audit.
tare
by fermionoid
Use this skill when the user explicitly asks to audit, prune, clean, or "tare" their accumulated agent scaffolds — phrases like "run tare", "audit my skills", "tare my scaffolds", "clean up my custom instructions", "my prompts feel bloated". Also appropriate to surface after a major model version has shipped, since past scaffolding tends to become overfit to past model behavior. Walks through the user's scaffolds (skill files, custom instructions, system prompts, slash commands, project rules), judges each one's value vs cost given current model capabilities, proposes KEEP / MODIFY / REMOVE for the user to confirm, then executes the change where the agent has file-system tools — otherwise outputs the exact change for the user to apply manually. Do NOT use for general coding tasks, single-shot prompts, debugging individual outputs, or anything other than explicit scaffold auditing.