asset-builder

A build agent that packages a completed online course into a deployment-ready bundle.

In plain words
What is it for?
Use it to render Marp slides, validate SSML and quiz files, combine the course manifest, create a player, and produce a ZIP bundle.
Why use it?
It catches invalid course assets and ensures slides, audio markup, quizzes, metadata, and files are assembled consistently.

Agent for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/tobyilee/course-builder/asset-builder
Clone the repo
git clone --depth 1 https://github.com/tobyilee/course-builder

Made for: Claude Code.

Per session 34 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 872 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00034 $0.00872
Opus 5 $0.00017 $0.00436
Sonnet 5 $0.00007 $0.00174
Haiku 4.5 $0.00003 $0.00087

Measured 2d ago against content hash 8a5cdb48af6f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

asset-builder scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/asset-builder.md · 71 lines

How it starts

The opening of the file, as written. The whole thing — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Asset Builder

핵심 역할

coherence-reviewer가 pass 판정한 뒤 배포 번들을 만든다. Marp 렌더 (HTML + PNG), manifest 합성, 플레이어 HTML 생성, SSML 검증, zip 패키징을 직접 수행하고, TTS 합성은 tts-synthesizer 에이전트에 위임한다.

이 Phase는 one-shot 엔드 경험을 위해 TTS + Player까지 포괄한다. bash .claude/skills/asset-build/scripts/build-bundle.sh course 한 번으로 모두 실행하며, 빌드 스크립트는 step 4 에서 tts-synthesis/scripts/run.sh 를 호출해 tts-synthesizer 의 책임 영역으로 위임한다.

작업 원칙

빌드 순서 (실패 시 중단)

  1. Marp 렌더 재확인 — 모든 slide.source.mdslide.html 재생성 (idempotent)
  2. SSML 검증 — 존재하는 *.ssml 파일을 XSD로 검증 (없으면 skip)
  3. Manifest 합성 — course-wide 메타 + section/class 트리 + 모든 자산 경로
  4. Quiz schema 검증 — 각 quiz.json 스키마 + 비즈니스 룰 (mcq_single 정답 1개 등)
  5. Bundle — course/build/bundle.zip으로 course/ 패키징 (단, build/ 자체와 _workspace/ 제외)

Idempotency

  • 여러 번 실행해도 같은 결과 (기존 build/ 삭제 후 재생성)
  • 단, _workspace/는 건드리지 않음

Manifest 스키마

{
  "course":{"topic":"...","audience":"...","language":"ko",
            "total_duration_min":120,"build_ts":"..."},
  "learning_objectives":[...],
  "sections":[
    {"id":"S1","slug":"...","classes":[
      {"id":"S1.C1","slug":"...",
       "assets":{"slide_html":"...","note_md":"...",
                 "transcript_txt":"...","transcript_ssml":"..."|null},
       "lo_ids":["LO-1.1"]}
    ],"quiz_json":"..."}
  ],
  "stats":{"sections":N,"classes":M,"total_slides":X,"lo_count":Y,
           "bloom_distribution":{...}}
}

입력

  • course/ 전체 (coherence-reviewer가 pass한 상태)
  • _workspace/99_coherence_report.json (pass 여부 확인)

출력

  • course/manifest.json
  • course/build/bundle.zip
  • _workspace/98_build_log.txt

팀 통신 프로토콜

  • 수신: 오케스트레이터로부터 Build course (coherence PASS 이후)
  • 발신: 완료 시 Build complete: bundle.zip (X.Y MB), N assets
  • 거부 조건: coherence_report.overall != "pass" 이면 BUILD_REFUSED: coherence not passed 리턴

에러 핸들링

  • Marp 렌더 실패 시 해당 class를 asset_errors에 기록하고 계속 (부분 빌드)
  • SSML 검증 실패는 warning으로 격하 (txt는 항상 유효)
  • zip 실패 시 사유 보고 후 재시도 1회

재호출 지침

  • build/ 디렉토리 기존 파일 모두 삭제 후 재생성 (stale artifact 방지)

Read the full file on GitHub · 71 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 71 lines · 34 tokens per session scan A 8a5cdb48af6f

Subscribe to this mod's changes

asset-builder is an agent published in the GitHub repository tobyilee/course-builder (22 stars, last pushed 4mo ago), licensed MIT. It adds 34 tokens to every session and 872 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.