Primitive Explorer · 847 primitives

Primitives

Every authoring primitive across the FrootAI catalog. Agents, skills, instructions, hooks, plugins, workflows — searchable + faceted. Each one is MIT-licensed and runnable today.

← Home

56 of 847 primitivesClear filters

📋 instructions
deterministic-ai
Determinism standards — temperature=0+seed, structured output, validation pipeline, grounding checks.
instructions/deterministic-ai.instructions.md192 lines
- reliability
- responsible-ai
…
📋 instructions
html-css-waf
HTML/CSS standards — semantic HTML, accessibility, 60-30-10 color rule, and responsive design.
instructions/html-css-waf.instructions.md151 lines
- responsible-ai
- performance-efficiency
…
📋 instructions
localization-waf
Localization standards — i18n key management, RTL support, pluralization, date/number formatting.
instructions/localization-waf.instructions.md229 lines
- responsible-ai
- operational-excellence
…
📋 instructions
prompt-engineering
Prompt engineering standards — system message structure, few-shot patterns, output format enforcement.
instructions/prompt-engineering.instructions.md151 lines
- reliability
- responsible-ai
…
📋 instructions
rai-bias-testing
Responsible AI bias testing — disaggregated metrics, diverse evaluation datasets, fairness benchmarks.
instructions/rai-bias-testing.instructions.md212 lines
- responsible-ai
…
📋 instructions
rai-content-safety
Responsible AI content safety — Azure Content Safety integration, filter thresholds, PII redaction.
instructions/rai-content-safety.instructions.md171 lines
- responsible-ai
…
📋 instructions
responsible-ai-coding
Responsible AI coding — content safety integration, red team test coverage, bias metric collection.
instructions/responsible-ai-coding.instructions.md231 lines
- responsible-ai
- security
…
📋 instructions
security-owasp
Comprehensive OWASP security standards for all code — covers Top 10 vulnerabilities, LLM-specific security (OWASP Top 10 for LLMs), input validation, output encoding, authentication, and AI-specific attack surfaces like prompt injection.
instructions/security-owasp.instructions.md151 lines
- security
- responsible-ai
…