You add a .npmignore with one line, test/, and npm stops reading your .gitignore, so the next npm publish ships .env.production. Your .dockerignore says *.log, but Docker matches that only at the root, so logs/ still goes into the image. This skill finds those and the other ways ignore rules quietly fail, and tells you how to fix each one.

Your agent runs a scanner on the project folder. It decides what each tool would really do: git (every .gitignore including nested and parent ones, .git/info/exclude, negation, anchoring, folder-only patterns, and the .git/index list of files git already tracks, which .gitignore never untracks), Docker (the .dockerignore at the build context root or the Dockerfile-specific one, root-only patterns, ** and ! re-includes, matched against every COPY and ADD), and npm (the "files" list, .npmignore replacing .gitignore, workspaces, always-included and never-included files). Secret files are recognized by their content, not only their name: an .env file is reported only when it holds a real secret value. It never runs git, docker or npm. It returns BLOCK, REVIEW, or NOT FLAGGED, with the file, line, reason, and a safer approach. Secrets are never printed.

What it checks

BLOCK (fix before shipping)

• G01: secret files git would commit or already tracks (.env with keys, private keys, .p12/.jks key stores, service-account JSON, .npmrc/.pypirc/.netrc tokens, Terraform state, Rails master.key)

• G02: secret files or the .git folder copied into a Docker image by COPY or ADD

• G03: secret files npm publish would include

REVIEW (check before release)

• G11: node_modules, virtualenvs, caches and build output not ignored

• G12: COPY . with no .dockerignore, or dependency folders copied from the host

• G13: .dockerignore patterns that match only at the build context root

• G14: ! lines that re-include secret file names

• G15: files over 50 MB, SQLite databases, database dumps and logs not ignored

• G16: logs, coverage output and old tarballs npm publish would include

• G17: .DS_Store, Thumbs.db, swap files, *.orig and IDE workspace files not ignored

• G18: secret files in test, fixture and examples folders (often throwaway test keys, so check them)

What's included

• The scanner (Python 3.9+, standard library only)

• A rules guide with the reason and safer approach for each rule

• A risky and a safe example project, and the real report the scanner produces on the risky one

Limits

It reads the files on disk. Your global git excludes file, build contexts set only in docker-compose or CI commands, and files created during a build are not followed. Folders such as node_modules and dist are listed but not searched inside. NOT FLAGGED is not a guarantee.

Contents

One ZIP file with 38 files: SKILL.md, the scanner in scripts/ (5 Python files), references/rules.md (every rule with its reason and safer approach), and examples/ (30 example files plus report.md, the real scanner output on the risky example).

How to use: This is an agent skill: a folder with a SKILL.md file and Python scripts that an AI coding agent such as Claude Code, Cursor, or Codex CLI can run. Unzip it into your agent's skills folder (for Claude Code: ~/.claude/skills/), then ask your agent to check your project. The scripts also run on their own with Python 3.9 or newer, with nothing to install. They read your files only: no network access, and nothing is sent anywhere.