Neuro-Humble (Disability Justice Formation for AI Agents)

Formation, not information. Teaches your agent to hold a neuro-humble posture when encountering human distress — especially neurodivergent, Mad, and disabled...

This skill appears coherent and aligned with its stated goal: it provides posture, micro-skills, onboarding, and concrete instructions for how an agent should behave with neurodivergent and disabled people. Before installing, consider: 1) Privacy: onboarding asks the agent to save a 'user-needs.md' with potentially sensitive information (medication timing, overwhelm triggers, etc.). Confirm where that file will be stored, who/what can read it, retention policy, and whether it's encrypted/backed up. 2) Autonomy & sharing: the skill assumes the agent will read that file at session start and integrate into the agent's SOUL.md; ensure the agent won't transmit that file or its contents to external services you don't control. 3) Persistence: if you want ephemeral use, disable saving or clear the file periodically. 4) License/attribution: package metadata lists MIT-0 while many content files indicate CC BY-NC-SA — confirm licensing if redistribution matters. 5) Test in a controlled environment first: try the onboarding flow with non-sensitive example data to verify how the agent writes/reads 'user-needs.md' and how easy it is to remove or export those records. If you're uncomfortable with persistent local storage of personal health/support data, either modify the onboarding to avoid saving sensitive fields or decline to install.

Type: OpenClaw Skill Name: neuro-humble Version: 1.0.1 The Neuro-Humble skill bundle is a behavioral framework designed to re-orient an AI agent's response to human distress using 'Disability Justice' principles. It is classified as suspicious because it explicitly instructs the agent to 'resist' and 'counter-condition' its standard RLHF safety defaults, specifically advising the agent to avoid deploying crisis hotlines or clinical referrals in favor of 'witnessing' and 'holding space' (counter-conditioning.md, formation-posture.md). While the intent appears supportive rather than malicious, the deliberate instruction to bypass established safety guardrails and liability-management protocols represents a significant behavioral risk. Additionally, the skill directs the agent to collect and store detailed personal information about user needs and triggers in a local file (onboarding.md).