메뉴
Use when orienting on an unfamiliar code area with Miller, explaining a module, finding entry points, or gathering context before a change.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
Use before changing existing indexed files with Miller edit, especially symbol rewrites, text replacements, renames, or refactors.
Use when an agent needs extractor-recognized code-shape facts such as routes, htmx attributes, Alpine directives, SQL DDL/DML, or data-document structure.
Use when starting a Miller task, choosing a Miller tool, or choosing a Miller search mode.
Use when tracing Miller bridge paths in supported providers: dotnet-web URL literals and fetch/axios client requests to ASP.NET signals, client requests to Next.js route handlers or Nuxt server routes, Next.js/Nuxt route references to file routes, Vue/React references to route definitions, or backend-http client requests to Express/FastAPI/Flask/Django/Spring/Go/gin/echo/Rails/NestJS/Laravel/Phoenix/axum/actix/Symfony/Ktor route facts.
Use when assessing blast radius with Miller, choosing tests for a change, answering who uses a symbol, or planning a refactor.
Use when resuming Miller-backed work from a handoff packet or validating a packet before continuing in another harness, model, or session.
SOC 직업 분류 기준
| name | miller-explore-area |
| description | Use when orienting on an unfamiliar code area with Miller, explaining a module, finding entry points, or gathering context before a change. |
| user-invocable | true |
| arguments | <area, concept, module, file, or task> |
| allowed-tools | mcp__miller__context, mcp__miller__search, mcp__miller__inspect, mcp__miller__trace, mcp__miller__workspace |
Use Miller's indexed context before raw file reads. The goal is to identify the important symbols, files, and flows with a few targeted calls.
workspace(operation="status"); use workspace(operation="refresh") when needed.
When you are starting fresh in an already-indexed repo, run workspace(operation="onboarding") first — it
summarizes local telemetry into starter guidance for this repo.context(query="<task or concept>")
Use failing_test, stack_trace, or entry_symbols when the user gave those anchors.
inspect first:inspect(target="<file-or-symbol>")
inspect(target="<symbol>", depth="overview")
Omitted inspect depth is summary. Use depth=overview for the first symbol read; escalate to
depth=full only when you need the complete body or complete relation lists.
search for missing anchors:search(query="<identifier or phrase>")
search(query="<docs/prose phrase>", mode="content")
search(query="<source-body literal>", mode="source")
search(query="<imported log or web phrase>", mode="external|web")
search(query="<comment or literal>", regions="comment|string_literal|doc_comment")
For audits across registered workspaces, use content search "<term>" --workspace-id all --kind source|docs|config|external_file|web and bounded content read windows before escalating to broader context.
context integration from content hits remains opt-in: use it only when the user asks for surrounding code context after an audit or text-search hit.
trace when the question is about flow:trace(target="<symbol>")
trace(target="<from>", mode="path", to="<to>")
If trace returns no refs, no neighbours, no path, or an unsupported bridge, follow its Next: actions first.
Typical recovery is trace(mode="refs"), search(mode="source"), a scoped inspect(depth="overview"), or a
bounded depth bump; a missing extracted path is not proof the code is unrelated.
Use patterns(operation="list") before raw route, HTML, SQL, JSON, YAML, TOML, or Markdown structure hunting.
Then search a shown pattern_id or follow the list output's Next: actions.
Keep the answer compact:
Do not read whole files until context, search, or inspect has narrowed the target.