メニュー
Comprehensive AI risk management, EU AI Act compliance, and LLM usage governance
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/Hack23/homepage --skill ai-governanceコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
GitHub Agentic Workflows (gh-aw) - markdown-based AI automation with 5-layer security, safe outputs, and Continuous AI patterns
gh-aw CLI usage, compilation, testing, debugging, add-wizard, and CI/CD practices for GitHub Agentic Workflows
Multi-agent coordination, orchestrator-worker patterns, /plan decomposition, and project coordination for GitHub Agentic Workflows
SOC 職業分類に基づく
SKILL.md を表示中
| name | AI Governance |
| description | Comprehensive AI risk management, EU AI Act compliance, and LLM usage governance |
| license | Apache-2.0 |
| version | 1.0 |
| author | Hack23 AB |
| tags | ["ai-governance","eu-ai-act","llm-security"] |
| category | security |
| frameworks | ["EU AI Act","ISO 27001:2022","NIST AI RMF"] |
| related_policies | ["AI_Policy.md"] |
Enforce comprehensive AI risk management and EU AI Act compliance, based on AI Policy.
Key Principle: "AI systems require systematic risk management and human oversight."
ai_system_classification:
unacceptable_risk:
prohibited_systems:
- social_scoring: government_or_private_social_credit
- real_time_biometric: public_space_surveillance_without_warrant
- subliminal_manipulation: ai_that_manipulates_behavior
- exploit_vulnerabilities: ai_targeting_vulnerable_groups
action: never_develop_deploy_or_use
high_risk_systems:
examples:
- recruitment_ai: candidate_screening_or_selection
- credit_scoring: loan_approval_algorithms
- law_enforcement: predictive_policing_tools
requirements:
- risk_assessment: comprehensive_before_deployment
- data_quality: high_quality_representative_datasets
prohibited_ai_practices:
- unclassified_systems: deploying_ai_without_risk_assessment
- no_human_oversight: fully_automated_critical_decisions
- biased_training_data: using_unrepresentative_datasets
- black_box_systems: unexplainable_ai_for_high_risk_decisions
- sensitive_data_training: training_ai_on_personal_data_without_consent
- bypassing_controls: using_personal_ai_for_business_without_approval
copilot_workflow:
development:
enable_copilot: vscode_github_copilot_extension
use_suggestions:
- code_completion: accept_for_boilerplate_code
- function_generation: review_logic_before_accepting
- test_scaffolding: verify_test_coverage_and_assertions
security_checks:
pre_commit:
- secret_scanning: gitleaks_pre_commit_hook
- code_review: manual_review_of_ai_suggestions
- license_check: verify_no_copyleft_violations
ci_cd_pipeline:
- sast: sonarcloud_static_analysis
- sca: dependabot_vulnerability_scanning
- test_coverage: minimum_80_percent_coverage
prohibited_inputs:
never_type:
- api_keys: aws_access_keys_database_passwords
- customer_data: email_addresses_names_personal_info
- proprietary_algorithms: trade_secret_business_logic
- production_configs: database_connection_strings
documentation:
- attribution: note_ai_assisted_code_in_comments
- review_notes: document_manual_changes_to_ai_suggestions
- lessons_learned: track_copilot_false_positives_for_training
ai_risk_assessment:
system_identification:
name: recruitment_candidate_screening_tool
classification: HIGH_RISK_per_eu_ai_act
trigger: involves_employment_decisions
risk_analysis:
potential_harms:
- discrimination: algorithm_may_have_gender_or_age_bias
- privacy: processes_personal_data_from_resumes
- transparency: candidates_unaware_of_ai_involvement
likelihood: MEDIUM
impact: HIGH
overall_risk: HIGH
risk_mitigation:
technical_measures:
- bias_testing: fairness_metrics_on_diverse_test_set
- explainability: shap_or_lime_for_decision_explanations
- data_quality: representative_training_dataset
organizational_measures:
- human_review: recruiter_reviews_all_ai_recommendations
- transparency: candidates_informed_of_ai_use
- appeal_process: candidates_can_request_human_review
- audit: annual_third_party_fairness_audit
compliance_obligations:
eu_ai_act:
- risk_management_system: implemented_and_documented
- data_governance: high_quality_training_data_verified
- technical_documentation: available_for_authorities
- conformity_assessment: third_party_audit_planned
gdpr:
- lawful_basis: legitimate_interest_assessment
- data_minimization: only_relevant_resume_data_processed
- automated_decision: human_review_prevents_art_22_violation
approval:
decision: APPROVED_WITH_CONDITIONS
approver: CEO
date: 2026_02_10
review_date: 2026_08_10_semi_annual
conditions:
- quarterly_bias_testing
- monthly_human_override_rate_monitoring
- annual_external_audit
❌ Non-Compliant:
incorrect_chatgpt_use:
prompt: "Review this customer database schema and suggest optimizations"
attached: production_database_dump_with_customer_emails
risk: exposing_personal_data_to_third_party_ai
✅ Corrected:
compliant_chatgpt_use:
data_anonymization:
- remove_personal_data: replace_emails_with_placeholders
- sanitize_schema: generic_field_names_only
- sample_data: synthetic_test_data_not_production
prompt: "Review this anonymized schema and suggest optimizations"
attached: sanitized_schema_with_no_real_data
verification:
- no_personal_data: verified_before_submission
- no_secrets: no_connection_strings_or_credentials
- documented: ai_consultation_logged_in_project_notes
Policies: AI Policy, Information Security
Skills: owasp-llm-security, data-classification, privacy-policy
Frameworks: EU AI Act, NIST AI RMF, ISO 42001 (draft)
