본문으로 건너뛰기
cccskills
무료GitHub에 공개

example-argument-substitution

Test harness for Claude Code skill argument substitution — demonstrates capture-block pre-declaration, XML tag referencing, unintentional variable corruption in code blocks, and correct placement of shell examples in reference files. Use when verifying substitution behavior before applying a pattern to other skills, testing how arguments flow from skill invocations, or understanding the pre-declaration and reference file pattern with greet/farewell/inspect actions.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

설치 방법 보기

포함된 파일(3)

  • SKILL.md7.0 KB
  • references/argument-substitution-reference.md5.9 KB
  • SKILL-GOALS.md1.1 KB

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

Argument Substitution Pattern — Example Skill

How to Use This Skill

This skill is a living test harness for argument substitution behavior. Read it, run it, observe what happened, and extend it.

Step 1 — Read this file. Do this before running. Note what you expect to see in each section.

Step 2 — Run with 0 arguments: /example-argument-substitution

Step 3 — Run with 10 arguments: /example-argument-substitution CANARY_A CANARY_B CANARY_C CANARY_D CANARY_E CANARY_F CANARY_G CANARY_H CANARY_I CANARY_J

Step 4 — Compare. Each section states what it expects. Check whether output matches.


How to Add a New Test

When you have a hypothesis about substitution behavior — test it here before applying it anywhere else.

  1. State your hypothesis — write exactly what you expect to appear in the output
  2. Add the pattern — put it in a new section below with an **Expected with 10 args:** and **Expected with 0 args:** annotation
  3. Run with 0 args — /example-argument-substitution
  4. Run with 10 CANARY args — /example-argument-substitution CANARY_A CANARY_B ...
  5. Observe — does the rendered output match your hypothesis exactly?
  6. Record the finding — if the hypothesis was wrong, correct it. If it involves literal $N syntax, record the verified fact in ./references/argument-substitution-reference.md (reference files are not substituted)
  7. Only then apply the pattern to other skills

Do not document any pattern as safe without completing all 7 steps.


Compare each section below between the two runs. Some sections are intentional substitution — they should show values. Others are unintentional — they show what corruption looks like.


Capture Block (intentional substitution — should show values)

All positional args captured into named XML tags. Everything else references these tags.

<arg0>$0</arg0> <arg1>$1</arg1> <arg2>$2</arg2> <arg3>$3</arg3> <arg4>$4</arg4> <arg5>$5</arg5> <arg6>$6</arg6> <arg7>$7</arg7> <arg8>$8</arg8> <arg9>$9</arg9> <all_args>$ARGUMENTS</all_args> <arg_by_index_0>$ARGUMENTS[0]</arg_by_index_0> <arg_by_index_1>$ARGUMENTS[1]</arg_by_index_1> <arg_by_index_2>$ARGUMENTS[2]</arg_by_index_2>

Expected with 10 args: each tag holds its CANARY value. Expected with 0 args: all tags empty.


Intentional Substitution in Prose (should show values)

These are correct uses of substitution — injecting argument values directly into skill output.

The skill was invoked with first argument: <arg0> The target is: <arg1> All arguments received: <all_args>

A skill instruction that uses the value directly:

Process the file <arg1> using mode <arg0>.

Expected with 10 args: prose shows CANARY_A, CANARY_B, full CANARY string. Expected with 0 args: blanks where the values would be — still correct for a 0-arg invocation.


Intentional Substitution Combined with Command Substitution

Command substitution runs at load time; argument substitution also runs at load time. Both happen before Claude reads the skill. This line combines both:

!echo "Loaded at $(date '+%Y-%m-%dT%H:%M:%S') — first arg is: $0"

Expected with 10 args: timestamp + CANARY_A on the same line. Expected with 0 args: timestamp + empty string after first arg is:.


Unintentional Substitution — Code Examples (shows corruption)

This section intentionally demonstrates what corruption looks like when you write shell code examples in SKILL.md body. The variables below are consumed at load time.

A bash function you might want to document:

check_file() {
    local path=$1
    local mode=$2
    [[ -f $1 ]] && chmod $2 $1
}

Expected with 10 args: $1 → CANARY_A, $2 → CANARY_B — code is corrupted. Expected with 0 args: $1 → empty, $2 → empty — also corrupted, differently.

Brace form is equally substituted:

process() {
    echo "arg: ${1}, mode: ${2}"
}

Expected with 10 args: ${1} → CANARY_A, ${2} → CANARY_B — brace form is NOT safe. Expected with 0 args: both render empty — still corrupted.

Awk field references in single quotes are also substituted:

awk '{print $5, $1}' file.txt

Expected with 10 args: $5 → CANARY_F, $1 → CANARY_A — awk example is broken. Expected with 0 args: both disappear — single quotes provide no protection.


The Correct Pattern — Pre-Declaration + Reference File

For prose and output strings: use substitution directly (as shown in the Intentional sections above).

For code examples containing shell variables: move them to a reference file. Reference files are NOT subject to substitution and can show literal $1, ${1}, $ARGUMENTS syntax safely.

The pre-declaration pattern for routing skills:

  1. Capture args into XML tags at the very top of SKILL.md (as in the Capture Block above)
  2. Use <tagname> throughout the rest of SKILL.md — never bare $N after the capture block
  3. Put all code examples with $N in references/*.md

Routing (uses <arg0> as action — correct pattern)

flowchart TD
    Start(["Read <arg0>"]) --> Q{arg0 value?}
    Q -->|"greet"| Greet["Say hello to <arg1>"]
    Q -->|"farewell"| Farewell["Say goodbye to <arg1>"]
    Q -->|"inspect"| Inspect["Show all captured values"]
    Q -->|"(empty)"| Help["Output usage line"]
    Q -->|"(anything else)"| Unknown["Output: Unknown action. Valid: greet, farewell, inspect"]

Actions

greet

Trigger: <arg0> is greet

Output:

Hello, <arg1>!
(invoked as: <all_args>)

If <arg1> is empty, substitute world.

farewell

Trigger: <arg0> is farewell

Output:

Goodbye, <arg1>. It was a pleasure.
(invoked as: <all_args>)

If <arg1> is empty, substitute friend.

inspect

Trigger: <arg0> is inspect

Output all captured values:

arg0              = <arg0>
arg1              = <arg1>
arg2              = <arg2>
arg3              = <arg3>
arg4              = <arg4>
arg5              = <arg5>
arg6              = <arg6>
arg7              = <arg7>
arg8              = <arg8>
arg9              = <arg9>
all_args          = <all_args>
arg_by_index_0    = <arg_by_index_0>
arg_by_index_1    = <arg_by_index_1>
arg_by_index_2    = <arg_by_index_2>

Full Reference

All substitution variables, pitfall table, and verified escape evidence:

./references/argument-substitution-reference.md

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