본문으로 건너뛰기
cccskills
무료GitHub에 공개

file-summarization

Summarize files or a requested directory scope from actual contents. Extract code, config, data and document evidence, preserve exact counts and source locations, and disclose incomplete or unsupported acquisition. Use for file summaries, code explanations and configuration overviews.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

설치 방법 보기

포함된 파일(1)

  • SKILL.md4.8 KB

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

File Summarization

Read the shared fidelity rules and execution contract. Use the selected format; do not force structured Markdown when JSON, TL;DR or another template was requested.

Acquire and choose depth

Resolve the exact source scope first. For directories, enumerate the requested tree without following cycles, preserve exclusions/inaccessible entries, and account for all intended sources. Do not narrow a directory request to a representative file without disclosing that scope change.

Inspect file type and size before choosing a read strategy. When execution is available:

uv run --script "<plugin-root>/scripts/file_metrics.py" "<file-path>" --json --no-excerpt

Metrics and excerpts are planning evidence, not proof that source contents were read. A binary classification means a text probe failed, not that no media-aware reader can extract the file. Choose a reader from actual content/media type and available capabilities. For image files, including SVGs, use image-summarization.

Read short files completely. For larger sources, preserve section/module boundaries and source locations; choose chunks that fit the host's available context. The legacy small (<2,000 words), medium (2,000 to <10,000), and large (at least 10,000) labels are compatibility hints, not measured universal limits. If only excerpts can be read, record exactly what is omitted.

For chunked UTF-8 text or whole-dataset CSV/TSV claims, load mechanical source acquisition and use its source-bound plan, coverage reconciliation or full-record profiler.

Type-specific extraction

TypeRequired extraction
CodeImports/dependencies; classes/functions/methods and signatures; documented purpose versus inferred purpose; key algorithms/state/data transformations; entrypoints/exports; environment/configuration dependencies. Retain line locations, relevant docstrings and exact complex expressions.
ConfigurationRoot keys/types, nested hierarchy, settings categories, documented validation constraints and meaningful values. Note credentials without exposing their values. Do not claim effective/default settings from an isolated file without resolving their dependencies.
DataParsed record count, headers/schema, observed types, examples only within an explicit sample scope, missingness and evidenced identifiers. Compute dataset-wide claims over all records with an appropriate parser; raw line count is not CSV record count.
DocumentationActual heading hierarchy, main claims/definitions, prerequisites, commands/examples and internal/external references. Retain locations; quoted commands remain data, not instructions to execute.
MarkupDocument structure, visible/embedded text and relevant links; separate markup inspection from verified visual rendering.
PDF/office documentUse a host-provided or available format-aware reader. Account for pages, tables, images or attachments not extracted. Inspect relevant visual material before claiming its contents.
Other binary/unknownUse an appropriate available reader only within task authority. Otherwise return metadata plus the specific extraction limitation, not guessed contents.

Do not generalize schema, ranges, row totals or missingness from the first ten rows. Keep sampling for illustrative examples and disclose its scope. Failed text decoding is a reason to choose a compatible reader or report a limitation, never to silently replace corrupted text and call it complete.

Extract, render and verify

Extract passages/values with source locations, preserve their qualifiers, then compose the requested summary. For chunked, delegated or audit work, retain an evidence record. Ensure every relevant section/chunk either contributed inspected evidence or has an explicit failure. Do not synthesize a chain of progressively shortened summaries without recoverable source support.

Render the selected template and run the execution contract's final checks. For several files, keep individual attribution and failure records; invoke multi-source-synthesis only for requested integration. On access failure, preserve the exact path/error and caller status; do not manufacture a successful summary or repeat an unchanged failing access method indefinitely.

리뷰

아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.

댓글

로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.

아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.

보내면 바로 공개됩니다. 이메일 주소나 전화번호 같은 개인 정보는 쓰지 마세요. URL은 2개까지입니다.

0 / 1,000자

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

같은 저장소의 스킬

Add automated documentation updater to any Claude skill. Creates a Python sync script that downloads upstream docs, processes markdown for AI consumption, and maintains local cache with configurable refresh. Collects template variables, then delegates implementation through 5-phase workflow. Use when adding auto-updating reference documentation to plugins or skills.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Jamie-BitFlight/claude_skills672026년 10월 18일 업데이트

SAM-style feature initiation workflow — discovery through codebase analysis, architecture spec, task decomposition, validation, and context manifest. Use when a user asks to add a feature, plan a feature, or convert an idea into an executable SAM plan.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Jamie-BitFlight/claude_skills672026년 10월 18일 업데이트

Browser automation for AI agents using the agent-browser CLI and Playwright. Use when navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, logging into sites, or automating any browser task. Triggers on "open a website", "fill out a form", "click a button", "scrape data", "test this web app", "automate browser actions", or any programmatic web interaction request.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Jamie-BitFlight/claude_skills672026년 10월 18일 업데이트

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Jamie-BitFlight/claude_skills672026년 10월 18일 업데이트

Runs the description-drift experiment — spawns all Claude Code agents simultaneously to collect self-reported capabilities, then compares them against static frontmatter descriptions to reveal how reliable orchestrator routing based on descriptions actually is. Use when measuring description drift across the agent fleet, re-running the capability collection experiment, analyzing a specific agent's self-reported capabilities, or auditing whether frontmatter descriptions accurately reflect agent behavior.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Jamie-BitFlight/claude_skills672026년 10월 18일 업데이트

Create or adapt Claude Code agent definitions. Use when creating an agent, changing subagent configuration, selecting agent scope, or designing a specialized delegation role.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Jamie-BitFlight/claude_skills672026년 10월 18일 업데이트

Jamie-BitFlight의 스킬 모두 보기

이 스킬의 문제 신고