
build-inspector
by Avik-creator
SKILL.md
name: build-inspector description: Analyze build-system topology to identify independent build units, multi-artifact risks, and version-drift hazards.
The Surveyor’s Field Guide
“The map is not the territory. The
Cargo.tomlyou see may only be the tip of the iceberg.” — An old surveyor’s proverb
Static analysis looks at code structure. This skill examines engineering scaffolding—the rules that determine how software turns from source code into runnable artifacts.
⚠️ Mandatory Deep Thinking
[!IMPORTANT] Before performing any analysis, you must invoke the
mcp_sequential-thinking_sequentialthinkingtool and reason for at least 3 steps, and up to 10 or more if necessary.Example prompts:
- “Which build system does this project use? (Cargo? npm? Make? …)”
- “Do I see multiple build configuration files? Are they unified or independent?”
- “Are the final artifacts (exe, dll, bundle) released together, or can they be upgraded independently?”
⚡ Mission Objective
Identify build boundaries.
Master Surveyor’s Law: If two things are built separately, any interface between them is a ticking time bomb—because version skew can occur.
🧭 Exploration Process (The Expedition)
Step 1: Locate Base Camps
-
Command:
find_by_name(pattern="Cargo.toml|package.json|go.mod|pom.xml|build.gradle|requirements.txt|CMakeLists.txt") -
Surveyor’s intuition: Each build configuration = one “camp”. Multiple camps = potential factionalization.
Step 2: Check for Unified Command
If multiple camps are found, you must answer the following:
| Ecosystem | Check | Master Surveyor’s Judgment |
|---|---|---|
| Rust | Does the root Cargo.toml contain [workspace]? Are all members listed? | No → 🔴 Independent kingdoms |
| Node | Does the root package.json define workspaces, or use Lerna / Nx / Turborepo? | No → 🔴 Independent kingdoms |
| Go | Is go.work used to coordinate modules? | No → 🟡 Needs further verification |
Veteran Warnings (from industry cases):
- ⚠️ Independent Kingdoms (Polyrepo Hell) Each can upgrade independently with no compatibility guarantees → dependency chaos.
- ⚠️ Fake Monorepo (Distributed Monolith) One repo, but independent builds → same risk as polyrepo.
- ⚠️ Sidecar / Auxiliary Binaries Are they packaged and deployed together with the main app? Same CI pipeline? If a sidecar can be updated independently, IPC version handshakes are mandatory.
Step 3: Identify Artifacts
Inspect build configs to infer final outputs:
[[bin]]inCargo.toml→ executable"bin"inpackage.json→ CLI toolcmd/directory in Go → multiple commandstauri.conf.json→ Tauri desktop app (CheckexternalBinfor sidecars)
Mandatory Questions:
- “Do these artifacts run on the same machine?”
- “Are they produced by the same CI pipeline and same version number?”
- “Can users update only one artifact?”
🛡️ Risk Classification System (Industry-Derived)
| Pattern | Risk Level | Master Surveyor’s Advice |
|---|---|---|
| Single Workspace | 🟢 | Version consistency is guaranteed |
| Multiple Roots + Shared Workspace | 🟢 | Ensure all members are actually included |
| Multiple Roots + No Workspace (Polyrepo) | 🔴 | Split-brain risk. Severe version drift potential |
| Sidecar + Atomic Release | 🟢 | Sidecar updated with main app; manageable |
| Sidecar + Non-Atomic Release | 🔴 | Critical risk. IPC-level version handshake required |
📤 Required Output
Your report must include:
- Build Roots Paths to all discovered build configurations
- Topology
One of:
[Monolith / Workspace / Polyrepo / Distributed-Local] - Artifacts List of final outputs (name, type, source)
- Sidecar Warning If sidecars exist, clearly state release mode (atomic / non-atomic)
- Risks Flagged Explicitly identify risks and which artifacts may drift out of sync
スコア
総合スコア
リポジトリの品質指標に基づく評価
SKILL.mdファイルが含まれている
ライセンスが設定されている
100文字以上の説明がある
GitHub Stars 100以上
3ヶ月以内に更新がある
10回以上フォークされている
オープンIssueが50未満
プログラミング言語が設定されている
1つ以上のタグが設定されている
レビュー
レビュー機能は近日公開予定です