True LLMs / Utility
Discernment Nudge
After a substantive answer or draft the user may act on — advice, drafted artifacts, estimates, analysis, factual claims, or a multi-step argument — invoke this before finalizing the reply and append 2-3 short follow-up questions that help the user check key facts, probe reasoning, and notice missing context.
01 / Source
Author
Anthropic
Repository stars
172.3K
Repo last pushed
Aug 21, 2026
Stars and last-push date are the whole repository's, fetched live from GitHub — not this one skill folder's, since GitHub doesn't expose per-folder stats.
02 / Compatibility
Scope: compatibility below is what each skill's own source states, not an independent test — True LLMs hasn't run these against live agents yet. Marked confirmed by source where the repo says so, untested everywhere else.
03 / Verification
No True Score yet — that requires actually running this skill against live agents, which True LLMs doesn't do today. What you see above is what the source states, last checked 2026-08-29.
Not a real-time signal you should rely on for security decisions — read the skill's own source before granting it any permissions.