Marking a Postman variable as "secret" only masks it on screen: export the environment and the API key is in the JSON as plain text. An OAuth token you fetched can sit in the collection's auth block, and a saved example response can still hold a session cookie, so the next export, fork or git commit ships them. This skill finds those and ten other mistakes in API client files before you share, commit or import them, and tells you how to fix each one.

Your agent runs a scanner on Postman collections (v2.0 and v2.1), Postman environment and globals exports, Insomnia exports (v4 JSON or YAML, v5 YAML) and Bruno .bru request and environment files. It parses each file to the exact line, treats {{variables}}, {{vault:name}}, {{ _.token }}, {% response %} tags and {{process.env.NAME}} as references rather than secrets, skips the <API Key> style placeholders that OpenAPI imports leave, respects Bruno's vars:secret list, and reads pre-request and test scripts with their comments and strings understood. It never sends a request or runs a script. 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)

• P01: tokens, passwords, client secrets and keys written in auth blocks (bearer, basic, API key, OAuth 2.0 and 1.0, AWS, digest, NTLM, JWT and others)

• P02: credentials in headers (Authorization, X-API-Key, session cookies), query parameters or user:password@ URLs

• P03: passwords and secrets in request bodies (raw JSON, form fields, GraphQL)

• P04: secret-named environment, globals and collection variables with real values, Bruno vars with values written in the file, and scripts that set a literal secret

• P05: live provider keys and private keys anywhere in the file (GitHub, GitLab, Slack, Stripe live, Google, OpenAI, Anthropic, SendGrid, npm, Hugging Face, Postman API keys)

REVIEW (check before release)

• P11: http:// to remote hosts in requests or base-URL variables

• P12: strictSSL false, rejectUnauthorized: false or NODE_TLS_REJECT_UNAUTHORIZED

• P13: eval() or new Function() in pre-request and test scripts

• P14: scripts that send the whole environment, cookies or a secret to a host the collection does not use (OAuth token requests are recognized)

• P15: requests hard-coded to a production host instead of {{baseUrl}}

• P16: saved example responses and cookie jars holding session cookies, JWTs or access tokens

• P17: console.log of secret variables or whole variable stores

• P18: Authorization headers kept on redirects to another host (followAuthorizationHeader, forwardAuthorizationHeader)

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 (Postman collection and environment, Insomnia export, Bruno request and environment), and the real report the scanner produces on the risky one

Limits

It reads the exported files as written. Values that live only inside the Postman, Insomnia or Bruno app (local values, Postman Vault, Bruno secrets, app-wide settings such as global SSL verification) are not in the files and are not checked. Bruno's YAML format, Postman v1 collections, HAR files and Newman reports are not read, and files over 20 MB are listed as not checked. NOT FLAGGED is not a guarantee.

Contents

One ZIP file with 21 files: SKILL.md, the scanner in scripts/ (6 Python files), references/rules.md (every rule with its reason and safer approach), and examples/ (12 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.