60
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill cin12211/orca-q --skill research-expert- SKILL.md7.9 KB
Overview
This skill is a specialized research expert for parallel information gathering with structured outputs and clear file-based reporting. It is designed to detect the appropriate research mode from task instructions and execute an efficient, evidence-driven search strategy. Use it to get concise summaries plus a complete markdown report saved to a filesystem path for downstream review.
How this skill works
The skill parses the research objective, detects one of three modes (Quick Verification, Focused Investigation, Deep Research), and builds a staged search plan that moves from broad queries to targeted deep dives and gap-filling. It evaluates sources by priority (primary, academic, professional, news, general web), extracts quotations, data, and contradictions, then writes a full report to a timestamped markdown file and returns a lightweight summary. Tool usage is budgeted by mode and the process includes quality checks, error handling, and domain-specific adaptations.
When to use it
- You need a fast factual confirmation from authoritative sources (Quick Verification).
- You must explore a specific aspect of a technical or domain topic with structured findings (Focused Investigation).
- You require a comprehensive, multi-source deep dive including contradictions and gaps (Deep Research).
- You want a reproducible report saved to a file for audits or team review.
- You need parallelized searches and efficient use of web tools for time-sensitive topics.
Best practices
- Provide a clear, specific research objective and preferred depth (quick/focused/deep).
- Include key terms, domains, and any must-have sources or sites to prioritize early.
- Accept the saved markdown file as the canonical full report; summaries are intentionally lightweight.
- Request follow-up queries that reference the saved filename if you want iterative updates.
- When researching technical topics, specify versions, code examples, or system constraints upfront.
Example use cases
- Verify a single factual claim about PostgreSQL behavior before deployment (Quick Verification).
- Investigate differences between open-source database editors for feature comparison (Focused Investigation).
- Produce a deep report on migration strategies from SQLite to PostgreSQL including caveats and performance data (Deep Research).
- Gather authoritative sources and code samples for implementing a Nuxt + Electron editor integration (Focused Investigation).
- Create a reproducible research file for regulatory or compliance review of DBMS choices (Deep Research).
FAQ
No. The full comprehensive report is written to a timestamped markdown file; chat returns a concise summary and the file path for retrieval.
How do you decide search depth?
Depth is determined by detected mode keywords or explicit user instruction; Quick uses 3–5 tool calls, Focused 5–10, Deep 10–15 with early termination triggers if objectives are met.