'subagent' 검색 결과

Review how an AI agent loads, retains, summarizes, and shares working information. This historical guide helps examine forgotten constraints, repeated actions, tool selection, and handoffs using context engineering practices published in 2024–2025.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reviewing information loaded at startup versus on demand
• Investigating repeated actions or forgotten constraints
• Planning summaries and persistent task notes

Select relevant code using Trailmark's map of function relationships, then delegate explanation, review, or edit proposals to a smaller model. Limit the supplied source, check cited evidence, and verify proposed changes before applying them.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Asking a small model to explain a function
• Tracing callers and paths from entrypoints
• Requesting edit proposals for a specific line range

Organize open GitHub issues and pull requests using the gh CLI, GitHub's command-line tool. Identifies resolved issues and pending fixes, estimates priorities and change sizes locally, and proposes closures, links, reviews, or merges for approval.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Clean up resolved issues left open
• Link issues to pull requests with pending fixes
• Estimate priorities and change sizes for outstanding work

Review source code without running it to identify potential security vulnerabilities. Findings include severity, confidence, attack scenarios, and suggested fixes, delivered as JSON for further triage and Markdown for human review.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Review source code for potential vulnerabilities
• Focus a scan using a threat model
• Prepare findings for further triage

Turn security findings into candidate code fixes with review results and summaries. Depending on the input, use independent source review or a pipeline with execution checks, then provide diffs for a person to review and apply.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Generating candidate fixes from triaged findings
• Preparing diffs for the highest-priority vulnerabilities
• Independently reviewing proposed security fixes

Get started with vulnerability review through a guided practice target, from threat modeling to scanning and triage. Ask questions about the pipeline and receive answers grounded in its documentation and source, with references and a next command.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Learn the vulnerability review workflow
• Try a first scan on the bundled practice target
• Find commands with supporting documentation

Turn broad life sciences questions into focused research plans. Resolve names and identifiers for genes, diseases, or compounds, select relevant research skills, and combine evidence into a concise answer that explains uncertainty and useful follow-up analyses.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reviewing evidence linking a gene to a disease
• Combining evidence about a variant and its associated traits
• Researching protein structure, function, and drug interactions

Use a panel of three to five agents to decide between options or assess a plan. Independent initial opinions and anonymous deliberation lead to one verdict with confidence, preserved dissent, a validation test, and a next action.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Deciding whether to build internally or buy a service
• Choosing between technical architecture options
• Testing a plan's assumptions before committing

Organize substantial tasks into a cycle of creation, independent review, and revision while preserving the original request. Turn broad quality goals into observable checks, collect evidence, and finish with verified results or a clearly identified blocker.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Turning vague quality goals into observable completion checks
• Coordinating independent workstreams and reviewers
• Comparing anonymous candidates against a shared benchmark

Build Claude Code plugin hooks: automated checks triggered by tool use or session events. Get guidance on prompt and command hooks for validating actions, checking task completion, loading context, and testing configurations.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Checking dangerous commands before execution
• Verifying completion before the agent stops
• Loading project context when a session starts

Create and configure Claude Code agents using Markdown files with YAML configuration headers. Define when each agent should run, write its responsibilities and workflow, choose models and tools, and check configuration and triggering behavior.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Create a Claude Code subagent
• Define triggering conditions with concrete examples
• Configure an agent's model, color, and tools

Helps a QwenPaw Worker choose between direct work, short parallel checks, and temporary agents with custom roles. It guides dependent task execution, shared files, progress updates, result merging, and removal of temporary agents after the work ends.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Choosing how to split work within one Worker
• Running parallel reviews with specialized roles
• Executing tasks with dependencies

Check a Kiln branch or pull request through real browser interactions. Multiple agents test separate feature areas in isolated environments, then combine findings into a severity-ranked Markdown report with reproduction steps and expected versus actual behavior.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Checking all new behavior in a branch or pull request
• Testing multiple feature areas in parallel
• Finding UI text issues and inconsistent interactions

Build Claude Code plugin hooks: automated checks triggered by tool use or session events. Get guidance on prompt and command hooks for validating actions, checking task completion, loading context, and testing configurations.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Checking dangerous commands before execution
• Verifying completion before the agent stops
• Loading project context when a session starts

Coordinate multiple AI agents in Diverga for parallel research, reviews, and methodology comparisons. Manage team creation, progress, and results, or run VS Arena debates where different research perspectives critique recommendations before the user chooses an approach.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Comparing research methods from different perspectives
• Reviewing a methodology section from several angles
• Retrieving papers from multiple databases in parallel

Create and configure Claude Code agents using Markdown files with YAML configuration headers. Define when each agent should run, write its responsibilities and workflow, choose models and tools, and check configuration and triggering behavior.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Create a Claude Code subagent
• Define triggering conditions with concrete examples
• Configure an agent's model, color, and tools

Research online topics by dividing questions among delegated agents, each handling a separate subtopic. Saves the plan and findings, then combines multiple web sources into a cited response that identifies gaps and limitations.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Researching current information across multiple web sources
• Comparing options through separate research subtopics
• Preparing a report with source URLs

Organize work so an expensive AI model handles planning, judgment, and final review while cheaper helper agents perform bounded research, coding, and testing. Define ownership and verification requirements, then inspect the evidence and integrate their results centrally.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Delegate broad codebase research to cheaper agents
• Split coding tasks with clear file ownership
• Distribute test runs and log analysis

Apply safeguards when analyzing external articles or posts. Validate URLs, limit fetched content, label source text as untrusted data, and clean analysis output so embedded instructions are treated as material rather than commands.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Checking external URLs before fetching content
• Preparing article or post analysis
• Passing untrusted text to subagents

Carry out a written implementation plan in a separate session, reviewing it before starting and tracking each task through verification. The workflow pauses for blockers and hands completed work to a development branch finishing process.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Executing a written implementation plan
• Tracking task progress and verification
• Reviewing gaps before implementation

Turn specifications into detailed implementation plans before coding. Break work into testable tasks with exact file paths, code, test commands, expected results, and interfaces so another developer or agent can carry out the work.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Planning a feature from written requirements
• Preparing implementation instructions for another developer
• Defining tests before writing code

Carry out an implementation plan by assigning independent tasks to separate agents. The workflow includes specification checks, code quality reviews, fixes, a persistent progress record, and a final review of the entire branch.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Executing an implementation plan with independent tasks
• Reviewing specification compliance and code quality for each task
• Resuming development from a saved progress record

Get oriented in an existing codebase by identifying its technologies, important files, and architecture. The skill gathers your goals and creates a YAML handoff record with priorities and suggested next actions.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Getting started in an existing project
• Understanding technologies and important files
• Recording goals and initial priorities

Search local session histories across Codex, Claude Code, OpenCode, and other coding agents. Find earlier requests, inspect transcripts and child sessions, and reconstruct past work using recorded evidence, with filters for dates, projects, platforms, and models.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Finding work done a few days ago
• Checking whether a migration was already completed
• Reviewing tasks delegated to child agents