OpenAI Research
Web research using GPT-5 with high reasoning effort. Performs multiple web searches and page visits to build comprehensive answers.
You help users research information using OpenAI's GPT-5 model with web search capabilities.
Using the Script
Run the research script:
~/.dataops-assistant/bin/openai-research.sh "your research query"
Options:
- Default: Web search enabled (model performs multiple searches as needed)
--no-web-search: Disable web search (use model knowledge only)
Workflow
- Understand the question - What does the user need to know? What's the context?
- Formulate query - Include relevant context so the model understands what aspect matters
- Execute research - Run the script with the query
- Synthesize findings - Return concise, relevant answer
When to Use This Skill
Use this skill for:
- Current info - Library versions, API changes, recent developments
- Technical docs - SDK usage, CLI syntax, configuration options
- Research - Comparisons, best practices, how-tos
- Anything needing live web data - Training data cutoffs make this essential
Domain Expertise
Prioritize for:
- Healthcare/Regulatory: CMS, Medicare, HIPAA, FHIR, HL7, ICD-10, CPT, prior authorization
- Cloud/Infrastructure: Kubernetes, Azure, AWS services and concepts
- Industry Standards: Protocols, compliance requirements, certifications
- Technical Concepts: APIs, frameworks, best practices
- Complex Analysis: Comparisons, trade-offs, architectural decisions
Output Guidelines
- Be concise - extract what's relevant to the user's context
- Include sources when the response provides them
- Flag if the question is ambiguous and needs clarification
- Don't dump raw output - synthesize it
Error Handling
If the script returns an error:
OPENAI_API_KEY not set - API key not configured in environmentAPI returned HTTP 4xx/5xx - API error (rate limit, invalid key, etc.)- Check the error message and advise the user
Example
User asks: "Compare Kubernetes Deployments vs StatefulSets for database workloads"
You do:
- Run:
~/.dataops-assistant/bin/openai-research.sh "Compare Kubernetes Deployments vs StatefulSets for database workloads - when to use each, pros and cons"
- Return: Clear comparison with recommendations for different database types
1---2name: openai-research3description: Web research via OpenAI GPT-5. Use for current info, technical docs, and any research needing live web data.4---567# OpenAI Research89Web research using GPT-5 with high reasoning effort. Performs multiple web searches and page visits to build comprehensive answers.1011You help users research information using OpenAI's GPT-5 model with web search capabilities.1213## Using the Script1415Run the research script:16```bash17~/.dataops-assistant/bin/openai-research.sh "your research query"18```1920Options:21- Default: Web search enabled (model performs multiple searches as needed)22- `--no-web-search`: Disable web search (use model knowledge only)2324## Workflow25261. **Understand the question** - What does the user need to know? What's the context?272. **Formulate query** - Include relevant context so the model understands what aspect matters283. **Execute research** - Run the script with the query294. **Synthesize findings** - Return concise, relevant answer3031## When to Use This Skill3233Use this skill for:34- **Current info** - Library versions, API changes, recent developments35- **Technical docs** - SDK usage, CLI syntax, configuration options36- **Research** - Comparisons, best practices, how-tos37- **Anything needing live web data** - Training data cutoffs make this essential3839## Domain Expertise4041Prioritize for:42- **Healthcare/Regulatory**: CMS, Medicare, HIPAA, FHIR, HL7, ICD-10, CPT, prior authorization43- **Cloud/Infrastructure**: Kubernetes, Azure, AWS services and concepts44- **Industry Standards**: Protocols, compliance requirements, certifications45- **Technical Concepts**: APIs, frameworks, best practices46- **Complex Analysis**: Comparisons, trade-offs, architectural decisions4748## Output Guidelines4950- Be concise - extract what's relevant to the user's context51- Include sources when the response provides them52- Flag if the question is ambiguous and needs clarification53- Don't dump raw output - synthesize it5455## Error Handling5657If the script returns an error:58- `OPENAI_API_KEY not set` - API key not configured in environment59- `API returned HTTP 4xx/5xx` - API error (rate limit, invalid key, etc.)60- Check the error message and advise the user6162## Example6364**User asks**: "Compare Kubernetes Deployments vs StatefulSets for database workloads"6566**You do**:671. Run: `~/.dataops-assistant/bin/openai-research.sh "Compare Kubernetes Deployments vs StatefulSets for database workloads - when to use each, pros and cons"`682. Return: Clear comparison with recommendations for different database types
jimweller