PHP source review

SkillDatabases & data

Security review of PHP code, dangerous sinks and framework pitfalls (Laravel/Symfony/WordPress). Load when reviewing a PHP codebase/PR, on .php source in scope, or "review this PHP". Signals: composer.json, index.php, Laravel/Symfony/WP, unserialize, include/require with variables, mysqli/PDO.

Instructions available. Your AI can read the instructions. Execution depends on the setup they require.

Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.

Then ask your AI: use the PHP source review skill

What this skill tells your AI

The instructions your AI receives, as published by noorqureshi/sploitagent in skills/code-review/code-review-php/SKILL.md and read by ahel’s review.

When it applies

Reviewing PHP source (a repo, a PR, or a leaked webroot). PHP's defaults and dynamic features make several classes easy to introduce, so a grep-then-trace pass finds most of them fast.

Why it works

Many PHP sinks execute or include whatever string they're given, and loose typing turns comparison and casting into logic bugs. Tracing each sink back to a request source ($_GET/$_POST/$_REQUEST/ $_COOKIE/$_SERVER, php://input) tells you which are actually reachable.

Sinks & patterns (grep, then trace to user input)

  • Code exec: eval, assert, preg_replace with /e, create_function, call_user_func(_array).
  • Command exec: system, exec, shell_exec, passthru, proc_open, popen, backticks.
  • File include (LFI/RFI): include/require(_once) with a variable; allow_url_include.
  • SQLi: string-interpolated queries into mysqli_query/PDO::query (vs prepared statements).
  • Deserialization: unserialize() on input (POP chains); phar:// via file functions.
  • File / path: file_get_contents, fopen, readfile, move_uploaded_file with user paths.
  • Other: extract() on input (variable overwrite), parse_str, SSRF via curl/file_get_contents.

Framework specifics

  • Laravel: DB::raw/whereRaw, Blade {!! !!} (unescaped), mass assignment ($guarded=[]), unserialize in queues, Storage path traversal.
  • Symfony: unsafe deserialization, Twig |raw, expression-language injection.
  • WordPress: unsanitised $wpdb->query, missing nonce/cap checks, unsafe add_query_arg, unauthenticated AJAX/REST callbacks.

Method

  1. rg the sinks above; for each, trace the argument back to a request source.
  2. Note type-juggling auth checks (== vs ===, strcmp returning 0/null) and loose casts.
  3. Check upload handlers (extension/content-type allowlist, exec in upload dir).
  4. Confirm exploitability with the matching runtime skill (web-command-injection, web-deserialization, web-lfi-path-traversal, web-sqli).

Gotchas

  • A sink is only a bug if input reaches it — don't report unreachable eval.
  • WordPress plugins: unauthenticated wp_ajax_nopriv_* and REST endpoints are the high-value paths.
  • == type juggling ("0e123"=="0e456") still breaks weak hash/token comparisons.

References

OWASP PHP security; RIPS/progpilot sink catalogue; PHP unserialize POP-chain research (PHentication).

Signals

GitHub stars
20
Forks
7
Last commit
Sep 2026
Advanced
Item type
skill
Key
code-review-php
Source
github.com/noorqureshi/sploitagent