スキル一覧に戻る
pentaxis93

technical-review

by pentaxis93

AI and I - a blog about human-AI collaboration

0🍴 0📅 2026年1月23日
GitHubで見るManusで実行

SKILL.md


Technical Review Skill

Purpose: Verify every technical claim before publication.


When to Use

At the TECHNICAL stage, after VOICE pass, before POLISH.


The Gate Question

Would you stake your professional reputation on every sentence?

If any sentence makes you hesitate, it needs verification.


Review Categories

1. Code Examples

Verify:

  • Code actually runs
  • Output matches what's shown
  • No syntax errors
  • Imports/dependencies noted
  • Version-specific behavior documented

Process:

# For each code example:
1. Extract to standalone file
2. Run it
3. Compare output to article
4. Note any environment requirements

Common issues:

  • Code worked when written, now fails (dependency updates)
  • Missing context (imports, setup)
  • Simplified for readability but now doesn't run

Verify:

  • All links resolve (no 404s)
  • Links go where expected (no redirects to wrong place)
  • External content still says what you're citing

Process:

# Extract all links, check each
curl -I [url]  # Check for 200 OK

Common issues:

  • Link rot (content moved or deleted)
  • Paywall appeared
  • Content changed since you read it

3. Technical Claims

Verify:

  • Tool/library does what you say it does
  • Version numbers are correct
  • Performance claims have evidence
  • "Best practice" claims have authoritative source

Questions to ask:

  • Where did I learn this? Is that source reliable?
  • Has this changed since I learned it?
  • Am I oversimplifying in a misleading way?

4. Dates & Numbers

Verify:

  • Event dates are accurate
  • Version numbers are correct
  • Statistics have sources
  • Timelines are consistent within article

Common issues:

  • Memory is wrong about when something happened
  • Mixing up version numbers
  • Outdated statistics

5. Names & Attribution

Verify:

  • Tool/library names spelled correctly
  • People's names spelled correctly
  • Correct attribution for quotes/ideas
  • License compliance for code snippets

Review Output Format

## Technical Review: [Article Title]

### Code Examples
| # | Description | Status | Notes |
|---|-------------|--------|-------|
| 1 | [desc] | PASS/FAIL | [issue if any] |

### Links
| Link | Status | Notes |
|------|--------|-------|
| [url] | PASS/FAIL | [issue if any] |

### Claims
| Claim | Verified | Source |
|-------|----------|--------|
| "[claim]" | YES/NO | [source or issue] |

### Dates & Numbers
| Item | Verified | Notes |
|------|----------|-------|
| [date/number] | YES/NO | [correction if needed] |

### Verdict: [PASS / NEEDS FIXES]

### Required Fixes:
1. [Specific fix]
2. [Specific fix]

Severity Levels

LevelDescriptionAction
CriticalFactually wrong, misleadingMust fix before publish
MajorBroken code, dead linksShould fix before publish
MinorTypo in name, outdated but not wrongFix if time permits

Anti-Patterns

Anti-PatternProblemFix
"I remember this working"Memory isn't reliableRun it again now
"The link was fine yesterday"Links breakCheck at review time
"Everyone knows this"Assumption may be wrongFind authoritative source
"Close enough"Erodes trustGet it exactly right

Trust Maintenance

Technical accuracy is trust infrastructure. One wrong fact can:

  • Undermine the entire article
  • Damage credibility for future articles
  • Mislead readers who trust you

The standard: Every technical claim must be verifiable.


Accuracy is not perfectionism. It's respect for the reader.

スコア

総合スコア

50/100

リポジトリの品質指標に基づく評価

SKILL.md

SKILL.mdファイルが含まれている

+20
LICENSE

ライセンスが設定されている

0/10
説明文

100文字以上の説明がある

0/10
人気

GitHub Stars 100以上

0/15
最近の活動

3ヶ月以内に更新がある

0/10
フォーク

10回以上フォークされている

0/5
Issue管理

オープンIssueが50未満

+5
言語

プログラミング言語が設定されている

+5
タグ

1つ以上のタグが設定されている

0/5

レビュー

💬

レビュー機能は近日公開予定です