ワンクリックでManusで任意のスキルを実行

techdebt

スター2

フォーク0

更新日2026年4月22日 22:14

Scan for technical debt — TODOs, obsolete APIs, EOL frameworks, nullable warnings suppressed, analyzer warnings silenced, deprecated NuGets. Produces ranked backlog.

インストール

Codex または Claude でインストールこの Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。

Manusで実行

ソース

dimsour

dimsour/dotnet-ai-toolkit

GitHub リポジトリを開く Creator のリポジトリを見る

ダウンロード

Manusで実行

techdebt

Surface accumulated debt with concrete remediation estimates.

What Counts as Tech Debt

// TODO, // HACK, // FIXME comments (grep + cite)
#pragma warning disable without a comment explaining why
[Obsolete] APIs still in use
Suppressed nullable warnings (! everywhere, #nullable disable)
NuGet packages ≥2 major versions behind
Target framework EOL (net6.0, netcoreapp3.1, etc.)
Deprecated EF patterns (Include strings, UseLazyLoadingProxies without intent)
Synchronous APIs where an async equivalent exists
Old-style Startup.cs on .NET 6+ (should be top-level Program.cs)
Test projects missing coverlet.collector or with <50% coverage
GlobalSuppressions.cs / .editorconfig rules set to none
TreatWarningsAsErrors=false in Release configuration

Flow

Scan via Grep / dotnet list package --outdated / .csproj property parsing
Categorize into: Framework, Dependencies, Code smells, Suppressions, Tests, Config
Rank by impact × ease:
- 🔴 Security-relevant (EOL framework, unpatched CVE) — top
- 🟠 Correctness risk (suppressed nullable in hot code) — mid-high
- 🟡 Maintenance friction (outdated but working) — mid
- 🟢 Nits — bottom
Estimate effort per item (S/M/L) with rationale
Output ranked list → optional /dotnet:plan handoff

Iron Laws

Fixing debt must not violate Iron Laws (don't suppress a nullable warning instead of fixing it)
Don't classify a business decision as debt (feature flag kept on purpose)
Don't suggest bulk upgrades without checking breaking-change notes

Output

.claude/audit/techdebt.md:

# Technical Debt — <date>

## Summary
- Critical: 3
- High: 7
- Medium: 14
- Low: 22

## Critical

### 1. Target framework net6.0 (EOL 2024-11)
- File: Directory.Build.props:3
- Effort: M (2–3 days — test project runner, ASP.NET Core APIs)
- Rationale: security patches stopped
- Fix: bump to net10.0 LTS (current LTS) or net8.0 if constrained

### 2. `Microsoft.AspNetCore.Authentication.JwtBearer` v5.0.17 — CVE-2023-...
- File: src/Api/Api.csproj:12
- Effort: S
- Fix: upgrade to ≥8.0.x

Integration

/dotnet:techdebt → .claude/audit/techdebt.md
        ↓
/dotnet:plan "Pay down Critical + High tech debt"
        ↓
/dotnet:work

References

${CLAUDE_SKILL_DIR}/references/detection-patterns.md — grep patterns for each debt category
${CLAUDE_SKILL_DIR}/references/estimation.md — S/M/L effort rubric with .NET-specific examples
${CLAUDE_SKILL_DIR}/references/upgrade-playbooks.md — common multi-step upgrades (net6→net8, EF6→EFCore)

Anti-patterns

Flagging every // TODO as critical — they're mostly Low
Suggesting upgrades without CVE/breaking-change context
Missing tests classified as Critical when the module is dead code
Bulk "upgrade everything" plan — always prioritize

このリポジトリの他の Skills

同じリポジトリ

audit

dimsour/dotnet-ai-toolkit

Full-project health audit — spawns 5 specialists (dotnet-reviewer, security-analyzer, testing-reviewer, performance-profiler, deployment-validator) in parallel. Use for onboarding or pre-release.

2026-04-222

boundaries

dimsour/dotnet-ai-toolkit

Analyze project/namespace boundaries, detect circular refs and layering violations via dotnet-depends and solution graph. Use before major refactors or to validate Clean Architecture.

2026-04-222

challenge

dimsour/dotnet-ai-toolkit

Senior-engineer challenge mode — aggressively question the current approach, hunt for N+1 queries, captive deps, missing auth, sync-over-async, leaks. Use before finalizing significant changes.

2026-04-222

compound

dimsour/dotnet-ai-toolkit

Capture a solved problem as institutional knowledge. Writes `.claude/solutions/{category}/{slug}.md` with symptoms, root cause, fix, category, tags. Use after notable fixes.

2026-04-222

document

dimsour/dotnet-ai-toolkit

Generate or update documentation — XML doc comments, README, architecture docs, OpenAPI descriptions. Uses code as source of truth; no hallucinated features.

2026-04-222

learn-from-fix

dimsour/dotnet-ai-toolkit

Extract durable lessons from a just-finished fix — propose CLAUDE.md rule, Iron Law addition, or Solution doc. Use after resolving a surprising/repeatable bug.

2026-04-222

techdebt

Surface accumulated debt with concrete remediation estimates.

What Counts as Tech Debt

// TODO, // HACK, // FIXME comments (grep + cite)

#pragma warning disable without a comment explaining why

[Obsolete] APIs still in use

Suppressed nullable warnings (! everywhere, #nullable disable)

NuGet packages ≥2 major versions behind

Target framework EOL (net6.0, netcoreapp3.1, etc.)

Deprecated EF patterns (Include strings, UseLazyLoadingProxies without intent)

Synchronous APIs where an async equivalent exists

Old-style Startup.cs on .NET 6+ (should be top-level Program.cs)

Test projects missing coverlet.collector or with <50% coverage

GlobalSuppressions.cs / .editorconfig rules set to none

TreatWarningsAsErrors=false in Release configuration

Flow

Scan via Grep / dotnet list package --outdated / .csproj property parsing

Categorize into: Framework, Dependencies, Code smells, Suppressions, Tests, Config

Rank by impact × ease:

🔴 Security-relevant (EOL framework, unpatched CVE) — top
🟠 Correctness risk (suppressed nullable in hot code) — mid-high
🟡 Maintenance friction (outdated but working) — mid
🟢 Nits — bottom

Estimate effort per item (S/M/L) with rationale

Output ranked list → optional /dotnet:plan handoff

Iron Laws

Fixing debt must not violate Iron Laws (don't suppress a nullable warning instead of fixing it)

Don't classify a business decision as debt (feature flag kept on purpose)

Don't suggest bulk upgrades without checking breaking-change notes

Output

.claude/audit/techdebt.md:

# Technical Debt — <date> ## Summary - Critical: 3 - High: 7 - Medium: 14 - Low: 22 ## Critical ### 1. Target framework net6.0 (EOL 2024-11) - File: Directory.Build.props:3 - Effort: M (2–3 days — test project runner, ASP.NET Core APIs) - Rationale: security patches stopped - Fix: bump to net10.0 LTS (current LTS) or net8.0 if constrained ### 2. `Microsoft.AspNetCore.Authentication.JwtBearer` v5.0.17 — CVE-2023-... - File: src/Api/Api.csproj:12 - Effort: S - Fix: upgrade to ≥8.0.x

Integration

/dotnet:techdebt → .claude/audit/techdebt.md ↓ /dotnet:plan "Pay down Critical + High tech debt" ↓ /dotnet:work

References

${CLAUDE_SKILL_DIR}/references/detection-patterns.md — grep patterns for each debt category

${CLAUDE_SKILL_DIR}/references/estimation.md — S/M/L effort rubric with .NET-specific examples

${CLAUDE_SKILL_DIR}/references/upgrade-playbooks.md — common multi-step upgrades (net6→net8, EF6→EFCore)

Anti-patterns

Flagging every // TODO as critical — they're mostly Low

Suggesting upgrades without CVE/breaking-change context

Missing tests classified as Critical when the module is dead code

Bulk "upgrade everything" plan — always prioritize

name	techdebt
description	Scan for technical debt — TODOs, obsolete APIs, EOL frameworks, nullable warnings suppressed, analyzer warnings silenced, deprecated NuGets. Produces ranked backlog.
argument-hint	<optional: scope path>
effort	medium