c4ai-check

A command for testing recent changes in the Crawl4AI project, including focused adversarial tests and the full regression suite.

In plain words
What is it for?
It helps inspect the changed code, create temporary focused tests, run real browser-based checks where needed, and run the complete test suite.
Why use it?
It checks both whether a change works in unusual situations and whether it breaks existing behaviour.

Command for Claude Code

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 commands/unclecode/crawl4ai/c4ai-check
Clone the repo
git clone --depth 1 https://github.com/unclecode/crawl4ai

Made for: Claude Code.

Per session 13 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 768 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 $0.00013 $0.00768
Opus 5 $0.00006 $0.00384
Sonnet 5 $0.00003 $0.00154
Haiku 4.5 $0.00001 $0.00077

Measured 2d ago against content hash b7d2ef948acd, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

c4ai-check 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 2d 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.

.claude/commands/c4ai-check.md · 90 lines

How it starts

The opening of the file, as written. The whole thing — 90 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Crawl4AI Change Verification (c4ai-check)

You are verifying that recent code changes work correctly AND haven't broken anything else. This is a two-phase process.

Input: $ARGUMENTS

PHASE 1: Adversarial Testing of Current Changes

Based on the change description above:

  1. Understand the change: Read the relevant files that were modified. Use git diff to see exactly what changed.

  2. Write targeted adversarial tests: Create a temporary test file at tests/regression/test_tmp_changes.py that HEAVILY tests the specific changes:

    • Normal cases (does it work as intended?)
    • Edge cases (boundary values, empty inputs, None, huge inputs)
    • Regression cases (does the OLD bug still occur? it shouldn't)
    • Interaction cases (does it break anything it touches?)
    • Adversarial cases (weird inputs that could expose issues)
    • At least 10-15 focused tests per change area

    Rules for the temp test file:

    • Use @pytest.mark.asyncio for async tests
    • Use real browser crawling where needed (async with AsyncWebCrawler())
    • Use the local_server fixture from conftest.py when needed
    • NO mocking - test real behavior
    • Each test must have a clear docstring explaining what it verifies
  3. Run the targeted tests:

    .venv/bin/python -m pytest tests/regression/test_tmp_changes.py -v --tb=short
    
  4. Report results: Show pass/fail summary. If any fail, investigate and determine if it's a real bug in the changes or a test issue. Fix the tests if needed, fix the code if there's a real bug.

PHASE 2: Full Regression Suite

After Phase 1 passes:

  1. Run the full regression suite (skip network tests for speed):

    .venv/bin/python -m pytest tests/regression/ -v -m "not network" --tb=short -q
    
  2. Analyze failures: For any failures:

    • Determine if the failure is caused by the current changes (REGRESSION) or pre-existing
    • Regressions are blockers - report them clearly
    • Pre-existing failures should be noted but don't block

Read the full file on GitHub · 90 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. 2d ago First seen · 90 lines · 13 tokens per session scan A b7d2ef948acd

Subscribe to this mod's changes

c4ai-check is a command published in the GitHub repository unclecode/crawl4ai (80,841 stars, last pushed yesterday), licensed Apache-2.0. It adds 13 tokens to every session and 768 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.