safety AI Agent Skills

safety-critical-patterns

196

athola

Guidelines from the NASA Power of 10 rules for writing robust, verifiable code. Adapted for general software development with context-appropriate rigor. Use when writing critical code paths, reviewing for robustness, improving code quality. Do not use as rigid requirements for all code - match rigor to consequence.

vibe-coding-guardian

45

kanyun-inc

Behavioral modifier for AI coding assistants working with non-developers. Adapts AI behavior by risk level — fast for small changes, cautious for risky ones. Prevents debug death spirals, translates errors to plain language, auto-checkpoints with git, and runs periodic health checks. Always active, zero manual trigger needed.

crisis-response-protocol

43

curiositech

Handle mental health crisis situations in AI coaching safely. Use when implementing crisis detection, safety protocols, emergency escalation, or suicide prevention features. Activates for crisis keywords, safety planning, hotline integration, and risk assessment.

mooco

30

SimHacker

Custom orchestrator design notes and skill interface expectations

smart-revert

15

oimiragieo

Git-aware smart revert for tracks, phases, and tasks. Handles rewritten history, finds related commits, and provides safe rollback with multiple confirmation gates.

target-safety

6

huifer

Target safety assessment including known toxicities, essentiality, and off-target concerns. Use for early safety risk evaluation and target selection. Keywords: target safety, toxicity, essential gene, knockout, safety risks

admet-prediction

6

huifer

ADMET (Absorption, Distribution, Metabolism, Excretion, Toxicity) prediction for drug candidates. Use for assessing drug-likeness, PK properties, and safety risks early in drug discovery. Keywords: ADMET, PK, toxicity, drug-likeness, DILI, hERG, bioavailability

Capability Evolver Pro

1

hanabi-jpn

Safe sandboxed self-evolution engine with 5-layer safety, automatic rollback, and evolution dashboard

Skill Guardian

1

hanabi-jpn

5-layer security scanner for AI agent skills — post-ClawHavoc protection