agentsclimarketplace
Sign in

Aie web security

Skill mrhinkle/aienterprise-skills/plugins/aie-web/skills/aie-web-security

Read-only defensive self-audit of a website or web app for common security gaps — exposed secrets in client code and git history, dependency vulnerabilities, missing or weak HTTP security headers (CSP, HSTS, X-Content-Type-Options, X-Frame-Options), HTTPS/TLS and mixed-content issues, unsafe DOM sinks (innerHTML/XSS), CSRF/cookie and CORS misconfigurations, and privacy-compliance gaps. Produces findings ranked by severity with plain-language fixes. This skill NEVER attacks a live target — it only reviews code, config, and public response headers you own. Trigger on "security audit," "security review," "security check," "is my site secure," "check for exposed API keys," "audit my security headers," "am I leaking secrets," "check my CSP," "HTTPS/SSL check," "XSS review," or "privacy compliance check."From its SKILL.md

Install
npx -y skills add mrhinkle/aienterprise-skills --skill aie-web-security

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

5.0 KB, 975 tokens by cl100k_base, as published. Nobody here has run it

Web Security Self-Audit (Defensive)

Runs a safe, read-only security review of a website or web app you control. It reviews source code, configuration, dependency manifests, git history, and public response headers. It does not scan, probe, fuzz, or attack anything — this is a self-audit, not a penetration test.

Scope and Safety Rules

  • Only audit assets the user owns or is explicitly authorized to review.
  • Read-only. Do not send crafted/malicious payloads, brute-force, or exploit anything.
  • Never print full secret values in output — mask them (re_****…abcd).
  • If asked to attack a third-party site or bypass a control, decline and redirect to the defensive scope.

Workflow

  1. Confirm scope — which repo/folder, which deployed URL (if any), and the stack (framework, host, package manager). Stay stack-agnostic; adapt commands to what you find.
  2. Run the checks in the order below, gathering evidence.
  3. Report using the severity format at the end.

Read reference/audit-checklist.md for the full itemized checklist across all areas. Read reference/security-headers.md for header values and host-specific config examples. Read reference/glossary.md if the user needs plain-language definitions.

Check Areas (summary)

  1. Secrets exposure (Critical) — Grep the working tree AND git history for API keys, tokens, passwords, private keys, connection strings. Confirm .env* and *.pem are git-ignored and not committed. Confirm secrets live in the host's environment-variable store, not in client-shipped bundles. Any secret prefixed for client exposure (e.g., build-time public vars) must NOT be a real secret. If a live secret is found, flag it Critical and instruct immediate rotation.
  2. Dependencies — Run the package manager's audit (npm audit, pnpm audit, yarn audit, pip-audit, etc.). Classify by severity, note fixable vs. breaking.
  3. Security headers — Check for Content-Security-Policy, Strict-Transport-Security (HSTS), X-Content-Type-Options, X-Frame-Options (or CSP frame-ancestors), Referrer-Policy, and Permissions-Policy. See reference/security-headers.md.
  4. HTTPS / TLS — Certificate validity, single 301 HTTP→HTTPS redirect, HSTS present, no mixed content (HTTP assets on an HTTPS page).
  5. Input handling / XSS — Flag innerHTML, dangerouslySetInnerHTML, document.write, and unescaped user input rendered to the DOM or into URLs. Recommend textContent and encodeURIComponent.
  6. CSRF / cookies — Confirm state-changing endpoints validate origin or use anti-CSRF tokens; confirm cookies set SameSite and, for sensitive cookies, HttpOnly + Secure.
  7. CORS — Flag wildcard Access-Control-Allow-Origin: * on any authenticated or sensitive endpoint.
  8. Third-party scripts — Inventory external scripts; recommend Subresource Integrity (SRI) and version pinning for CDN resources; note what data analytics/trackers collect.
  9. Abuse prevention — Confirm public form/serverless endpoints have spam protection (honeypot and/or rate limiting) and reasonable function timeouts.
  10. Privacy compliance — Confirm a linked privacy policy, disclosure of data collected and third parties used, cookie consent where required, and a contact path for data requests (GDPR/CCPA baseline).

Output Format

Lead with an overall grade and counts, then list findings.

Overall: B — 1 critical, 3 warnings, 4 recommendations

For each finding:

  • Severity — CRITICAL (fix before launch) / WARNING (fix within a week) / RECOMMENDATION (best practice)
  • What — the issue, with file path and line or the header that's missing
  • Why it matters — one plain-language sentence
  • Fix — exact code, config, or step

Order findings Critical → Warning → Recommendation.

References

What ships with it: 3 files

8.6 KB alongside SKILL.md

reference/

Gives 2 of the 12 instructions most audit compliance skills give in 975 tokens

Counted across 937 of the 1,487 authors here whose files we hold, read 2026-08-07

  • Fetch latest guidelines before each reviewin 43 of 937, across 3 files
  • Group findings by severityhere, and in 43 of 937
  • Check files against all fetched rulesin 42 of 937, across 2 files
  • Output findings in terse file:line formatin 41 of 937, across 3 files
  • Ask user which files to review if none specifiedin 41 of 937, across 3 files
  • Read specified files or prompt user for filesin 39 of 937, across 1 file
  • Generate the audit reportin 33 of 937, across 30 files
  • Assign a severity to every findingin 25 of 937
  • Run automated accessibility scansin 23 of 937, across 13 files
  • Output a markdown audit reportin 22 of 937
  • Map findings to WCAG criteriain 20 of 937, across 10 files
  • Confirm audit scopehere, and in 19 of 937, across 9 files

Said here and by no other author read

  • audit only owned or authorized assets
  • review source code, config, manifests, git history, and headers
  • mask secret values in the output
  • decline requests to attack third-party sites
  • grep working tree and git history for secrets
  • check for missing HTTP security headers

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 326,835. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.