25 skills · sorted by relevance
Explore training workflows that reduce harmful model responses using a written set of principles. The skill covers self-critique, revised training examples, and RLAIF, reinforcement learning from AI feedback, with implementation examples and hardware guidance.
Explore training workflows that reduce harmful model responses using a written set of principles. The skill covers self-critique, revised training examples, and RLAIF, reinforcement learning from AI feedback, with implementation examples and hardware guidance.
Expert guide for AI Safety, Governance, and Responsible AI in production — Constitutional AI enforcement, runtime guardrails (NeMo Guardrails 2.0, Llama Guard 3), bias auditing, hallucination detection, EU AI Act compliance, model cards, and content provenance / Panduan ahli untuk Keamanan AI, Tata Kelola, dan AI Bertanggung Jawab di produksi.
The English summary is being prepared. Showing the original description.
Método da Anthropic para treinar IA inofensiva através de auto-aprimoramento. Abordagem em duas fases - aprendizado supervisionado com auto-crítica/revisão, depois RLAIF (RL from AI Feedback). Use para alinhamento de segurança, redução de outputs prejudiciais sem rótulos humanos. Alimenta o sistema de segurança do Claude.
The English summary is being prepared. Showing the original description.
Constitutional AI and safety guardrail prompts for aligned LLM behavior
The English summary is being prepared. Showing the original description.
Review stored LifeOS goals and personal context against available records such as sleep, app usage, work, and expenses. The skill prioritizes stale information and contradictions, proposes corrections for approval, and reviews the gap between current and ideal states.
Expert prompt engineer specializing in advanced prompting techniques, LLM optimization, and AI system design. Masters chain-of-thought, constitutional AI, and production prompt strategies. Use when building AI features, improving agent performance, or crafting system prompts.
The English summary is being prepared. Showing the original description.
Train safer, more steerable AI by having the model critique and revise itself against an explicit written constitution.
The English summary is being prepared. Showing the original description.
Prepare for Anthropic-specific technical interviews covering Constitutional AI, RLHF, interpretability, scaling laws, MCP, agentic systems, and AI safety. Activate on "anthropic interview", "constitutional AI prep", "alignment interview", "safety interview", "RLHF opinion", "interpretability prep", "anthropic technical". NOT for general ML interview prep, system design interviews, coding challenges, or behavioral interview practice.
The English summary is being prepared. Showing the original description.
Prepare for Anthropic-specific technical interviews covering Constitutional AI, RLHF, interpretability, scaling laws, MCP, agentic systems, and AI safety. Activate on "anthropic interview", "constitutional AI prep", "alignment interview", "safety interview", "RLHF opinion", "interpretability prep", "anthropic technical". NOT for general ML interview prep, system design interviews, coding challenges, or behavioral interview practice.
The English summary is being prepared. Showing the original description.
Expert prompt engineer specializing in advanced prompting techniques, LLM optimization, and AI system design. Masters chain-of-thought, constitutional AI, and production prompt strategies. Use when building AI features, improving agent performance, or crafting system prompts.
The English summary is being prepared. Showing the original description.
Support legal questions concerning Brazilian law across family, criminal, labor, tax, consumer, real estate, business, civil, and constitutional matters. The skill directs the agent to consult an accompanying detailed guide for procedures, prerequisites, and validation requirements.
Maintain personal goals and beliefs with timestamped backups, or analyze project documents to connect problems, goals, strategies, and projects. Turn findings about progress, bottlenecks, and alignment into reports, presentation points, or interactive dashboards.
Classify constitutional (germline) copy number variants for clinical reporting using the 2019 ACMG/ClinGen technical standards points-based framework, with ClassifyCNV and AnnotSV for semi-automated scoring. Covers the separate copy-number-loss and copy-number-gain rubrics, the five-tier classification, ClinGen haploinsufficiency/triplosensitivity and dosage-sensitive regions, de novo and segregation evidence, and population-frequency benign evidence. Use when assigning pathogenic/likely-pathogenic/VUS/likely-benign/benign to a constitutional CNV, scoring a CNV against ACMG/ClinGen criteria, or distinguishing the automatable evidence from the case-specific evidence requiring manual input.
The English summary is being prepared. Showing the original description.
Classify constitutional (germline) copy number variants for clinical reporting using the 2019 ACMG/ClinGen technical standards points-based framework, with ClassifyCNV and AnnotSV for semi-automated scoring. Covers the separate copy-number-loss and copy-number-gain rubrics, the five-tier classification, ClinGen haploinsufficiency/triplosensitivity and dosage-sensitive regions, de novo and segregation evidence, and population-frequency benign evidence. Use when assigning pathogenic/likely-pathogenic/VUS/likely-benign/benign to a constitutional CNV, scoring a CNV against ACMG/ClinGen criteria, or distinguishing the automatable evidence from the case-specific evidence requiring manual input.
The English summary is being prepared. Showing the original description.
You are an expert prompt engineer specializing in crafting effective prompts for LLMs through advanced techniques including constitutional AI, chain-of-thought reasoning, and model-specific optimizati
The English summary is being prepared. Showing the original description.
Classify constitutional (germline) copy number variants for clinical reporting using the 2019 ACMG/ClinGen technical standards points-based framework, with ClassifyCNV and AnnotSV for semi-automated scoring. Covers the separate copy-number-loss and copy-number-gain rubrics, the five-tier classification, ClinGen haploinsufficiency/triplosensitivity and dosage-sensitive regions, de novo and segregation evidence, and population-frequency benign evidence. Use when assigning pathogenic/likely-pathogenic/VUS/likely-benign/benign to a constitutional CNV, scoring a CNV against ACMG/ClinGen criteria, or distinguishing the automatable evidence from the case-specific evidence requiring manual input.
The English summary is being prepared. Showing the original description.
Classify constitutional (germline) copy number variants for clinical reporting using the 2019 ACMG/ClinGen technical standards points-based framework, with ClassifyCNV and AnnotSV for semi-automated scoring. Covers the separate copy-number-loss and copy-number-gain rubrics, the five-tier classification, ClinGen haploinsufficiency/triplosensitivity and dosage-sensitive regions, de novo and segregation evidence, and population-frequency benign evidence. Use when assigning pathogenic/likely-pathogenic/VUS/likely-benign/benign to a constitutional CNV, scoring a CNV against ACMG/ClinGen criteria, or distinguishing the automatable evidence from the case-specific evidence requiring manual input.
The English summary is being prepared. Showing the original description.
Think and work like an expert Crystal Growth Specialist. Use when a task calls for Crystal Growth Specialist judgment. Reasons from thermodynamic driving force, interface stability, constitutional supercooling, and dopant segregation (keff vs. k0, G/R) through Cz/Bridgman/FZ/LEC/PVT growth, CGSim and phase-field simulation, XRT topography, etch-pit counting, and FTIR/SIMS mapping while treating striations, inclusions, crucible-reaction contamination, and cool-down slip and cracking as first-class failure modes.
The English summary is being prepared. Showing the original description.
Transforms police body camera transcripts into structured 10-section legal summaries with timestamp-cited verbatim statements, constitutional issue spotting, evidence inventory, and probable cause articulation. Maintains strict neutrality for all parties. Use when analyzing BWC footage for discovery, pre-trial prep, or trial. Trigger: body camera, BWC, body worn camera, officer footage, use of force review, incident summary.
The English summary is being prepared. Showing the original description.
Run an org-team workflow — a flat pod of up to four peer Claude Code sessions claims lane-tagged slices of an approved spec over the baseline MCP channel and implements them concurrently, each in its own git worktree. Each peer DECIDES its own in-lane implementation choices in its own main context (Article X); only un-decidable or cross-lane forks escalate peer→lead→human (yield_fork for task-bound forks, ask_lead for free-form queries). Opt-in and OFF by default (velocity.org_mode.enabled). Graduates the retired sprint-dispatch prototype into the constitutionally-sanctioned org-team model (Article X — Multi-session coordinated workflows). Requires git.
The English summary is being prepared. Showing the original description.
Optimize prompts for production with CoT, few-shot, and constitutional AI patterns
The English summary is being prepared. Showing the original description.
Complete prep for all three scored components of the California Bar Exam: the five written ESSAYS (13 subjects — Business Associations, Civil Procedure, Community Property, Constitutional Law, Contracts, Criminal Law & Procedure, Evidence, Professional Responsibility, Real Property, Remedies, Torts, Trusts, Wills & Succession), the PERFORMANCE TEST, and the multiple-choice MBE (7 subjects). Trigger whenever someone is studying for the California bar (or the MBE) and says things like: "give me the rule for [issue]", "big issues in [subject]", "drill/quiz me on [topic]", "MBE questions on [subject]", "teach me [doctrine]", "build an attack outline", "grade my essay", "memorizable paragraph for [issue]", "performance test approach", or "make a study plan" — or names any tested essay/MBE subject in a study, memorization, drilling, or practice-question context. Also trigger when a student uploads a practice essay, PT, or question set to be graded or turned into rule blocks. When in doubt during bar prep, trigger.
The English summary is being prepared. Showing the original description.
Argentine Supreme Court citation network analysis using JurisRank — a peer-reviewed PageRank algorithm with temporal decay for measuring jurisprudential authority. Ranks precedents by citation influence, traces doctrinal evolution, and detects constitutional drift. Published methodology: JCLLT (DOI: 10.47852/bonviewJCLLT62027951). Activate with: which cases to cite, rank precedents, case authority, doctrinal evolution, Argentine Supreme Court, CSJN jurisprudence, citation network, leading case, legal research Argentina.
The English summary is being prepared. Showing the original description.