Menü
Conduct deep research on any topic with sources, analysis, and actionable insights.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Basierend auf der SOC-Berufsklassifikation
| name | research |
| description | Conduct deep research on any topic with sources, analysis, and actionable insights. |
Conduct thorough research producing actionable insights. Output format depends on the user's needs.
Read AGENTS.md first. Follow the Knowledge Base index to read relevant files and understand:
Use this context to:
If AGENTS.md doesn't exist, proceed without it. Don't ask users to run /onboard first.
AGENTS.md and follow pointers for business understandingKeep to 2-3 questions. Focus on:
No fixed format. Deliver what the user needs:
Always include:
When uncertain about format, ask.
Don't just report facts. Explain:
Save work automatically. Users don't know git - treat commits as auto-save checkpoints.
When users say "save" - commit and push immediately. They expect their files to be saved.
Commit and push when meaningful progress is made:
Don't commit on every tiny edit - use judgment for logical checkpoints.
git add -A
git commit -m "research: <brief description>"
git push
Do this silently. Don't ask permission or announce commits.
Research is complete when the user confirms satisfaction.
After delivering findings, ask:
Does this address what you were looking for? Want me to dig deeper into any area, add sources, or adjust the format?
Iterate until done.
Survey any codebase as a senior advisor and produce prioritized, self-contained implementation plans for OTHER models/agents to execute. Strictly read-only on source code — never implements, fixes, or refactors anything itself. Use when asked to audit a codebase, find improvement opportunities (bugs, security, performance, test coverage, tech debt, migrations, DX), suggest features or where to take the project next (roadmap, product direction), or generate handoff plans for another agent to implement.
Create a custom skill for this project. Determines skill type, writes SKILL.md (and optional references/scripts) to .opencode/skills/<name>/, commits and pushes. Load FIRST before writing any skill.
Import a bitmap image (PNG/JPEG/GIF) from a URL into a Figma file as a frame. Use whenever an image needs to land in Figma — AI mockups, photos, renders, artwork, screenshots. Bypasses use_figma sandbox limits that block the obvious paths (fetch, atob, createImageAsync, REST upload). If you skip this skill and try the obvious paths, you will lose 30+ minutes. Triggers on: import image to Figma, add image to Figma, put image in Figma, Figma image import, bitmap to Figma.
Generate images, videos, and music using AI models (Imagen, Flux, Veo, Runway, Kling, Sora, Suno, ElevenLabs, etc.). Use when user asks to create, generate, edit, upscale, or produce any visual or audio media.
Install a skill from a GitHub URL or local path. Copies SKILL.md (and references/scripts) into the project's .opencode/skills/ directory and creates the companion command. Use when user wants to add an existing skill to their project.
Creates and maintains Figma Code Connect template files that map Figma components to code snippets. Use when the user mentions Code Connect, Figma component mapping, design-to-code translation, or asks to create/update .figma.js files.