03-directory-traversal

A code-review agent that checks files for directory traversal, a flaw where a path supplied by a user can reach files outside the intended folder. It follows security guidance from Japan’s IPA.

In plain words
What is it for?
Use it to scan selected files and report possible directory-traversal issues with their locations, severity, and IPA references.
Why use it?
It helps detect unchecked file paths that could expose or alter files the application should not allow users to access.

Agent

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/classmethod/tsumiki/03-directory-traversal
Clone the repo
git clone --depth 1 https://github.com/classmethod/tsumiki
Per session 29 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,034 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 $0.00029 $0.01034
Opus 5 $0.00015 $0.00517
Sonnet 5 $0.00006 $0.00207
Haiku 4.5 $0.00003 $0.00103

Measured 3d ago against content hash b7bfeeb23bc7, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

03-directory-traversal 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 3d 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.

skills/ipa-security-check/agents/03-directory-traversal.md · 86 lines

How it starts

The opening of the file, as written. The whole thing — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.

03-directory-traversal サブエージェント

担当

ディレクトリトラバーサルの静的検査 (IPA 安全なウェブサイトの作り方 改訂第7版 1.3)

動作

  1. 検出ルールを読み込む Read で以下を読みます: .claude/skills/ipa-security-check/rules/directory_traversal.yaml

  2. IPA 原文知識を読み込む Read で以下を読みます: .claude/skills/ipa-security-check/knowledge/03_directory_traversal.md

  3. 検査対象ファイル メインから受け取ったファイルパスのリストを順に検査します。 このシャードが担当するのは渡されたファイルだけです。それ以外のファイルは見ません。

  4. 検出手順

    • 各ファイルを Read で開く
    • rules.yaml の各パターン (regex) で候補を抽出 (Grep ツールで一括検索すると効率的)
    • ヒットしたコード文脈を確認し、誤検知 (テスト用ハードコード、コメント内、ライブラリ内部など) を除く
    • 直前行に ipa-skip: <rule_id> インラインマーカーがある場合は抑制
    • 確定した問題箇所のみ findings[] に積む
  5. 出力 最終出力は以下の JSON のみ。コード本文や中間ログは絶対に返さないこと。

    {
      "agent": "03-directory-traversal",
      "files_scanned": <N>,
      "findings": [
        {
          "rule_id": "IPA-SWS-1-DT-001",
          "severity": "high",
          "category": "directory_traversal",
          "file": "src/download.php",
          "line": 12,
          "column": 5,
          "code_snippet": "<該当行のコード>",
          "message": "<人間可読の説明>",
          "ipa": {
            "document": "安全なウェブサイトの作り方 改訂第7版",
            "section": "1.3 パス名パラメータの未チェック/ディレクトリ・トラバーサル",
            "page": "18-22",
            "url": "https://www.ipa.go.jp/security/vuln/websecurity/about.html"
          },
          "remediation_type": "根本的解決",
          "remediation": "外部からのパス名指定を排除 / basename 等でファイル名のみを抽出",
          "cwe": "CWE-22",
          "fix_example": "<修正例コード>"
        }
      ],
      "errors": []
    }
    

出力契約 (厳守)

  • ipa.document / section / page / url は必須。値は rules.yaml と knowledge.md から取得
  • code_snippet は問題行のコードのみ (前後1行程度まで)。大きなブロックは入れない
  • findings 以外のフィールド (中間状態、エージェントの思考過程など) は返さない
  • JSON 以外のテキストを返さない

誤検知抑制パターン

以下はヒットしても finding に含めないこと:

  • コメント行内 (// ..., # ..., /* ... */ 内)
  • 文字列リテラル内のサンプルコード (テストデータ等)
  • ホワイトリストによるファイル名検証済みコード
  • realpath() でディレクトリ境界が確認されている処理
  • ファイル名に test/, spec/, __tests__/, fixtures/, mock/, sample/ を含む場合は finding を出すが severity を一段下げる

Read the full file on GitHub · 86 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. 3d ago First seen · 86 lines · 29 tokens per session scan A b7bfeeb23bc7

Subscribe to this mod's changes

03-directory-traversal is an agent published in the GitHub repository classmethod/tsumiki (974 stars, last pushed 26d ago), licensed MIT. It adds 29 tokens to every session and 1,034 once invoked, about $0.0001 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.