Agent skill

implementing-fuzz-testing-in-cicd-with-aflplusplus

Integrate AFL++ coverage-guided fuzz testing into CI/CD pipelines to discover memory corruption, input handling, and logic vulnerabilities in C/C++ and compiled applications.

Stars 4,300
Forks 470

Install this agent skill to your Project

npx add-skill https://github.com/mukul975/Anthropic-Cybersecurity-Skills/tree/main/skills/implementing-fuzz-testing-in-cicd-with-aflplusplus

SKILL.md

Implementing Fuzz Testing in CI/CD with AFL++

Overview

AFL++ (American Fuzzy Lop Plus Plus) is a community-maintained fork of AFL that provides state-of-the-art coverage-guided fuzz testing for discovering vulnerabilities in compiled applications. AFL++ uses genetic algorithms to mutate inputs, tracking code coverage to find new execution paths that trigger crashes, hangs, and undefined behavior. In CI/CD environments, AFL++ can be integrated to continuously test parsers, protocol handlers, file format processors, and any code that handles untrusted input. AFL++ supports persistent mode for high-speed fuzzing (up to 100,000+ executions per second), custom mutators, QEMU mode for binary-only fuzzing, and CmpLog/RedQueen for automatic dictionary extraction.

When to Use

  • When deploying or configuring implementing fuzz testing in cicd with aflplusplus capabilities in your environment
  • When establishing security controls aligned to compliance requirements
  • When building or improving security architecture for this domain
  • When conducting security assessments that require this implementation

Prerequisites

  • Linux-based CI runners (AFL++ does not support Windows natively)
  • GCC or Clang compiler toolchain
  • AFL++ installed (apt install aflplusplus or built from source)
  • Target application with harness functions isolating input processing
  • Seed corpus of valid input samples

Core Concepts

Coverage-Guided Fuzzing

AFL++ instruments the target binary at compile time (or via QEMU/Frida for binary-only targets) to track which code paths each input exercises. When a mutated input triggers a new code path, it is saved to the corpus for further mutation. This feedback loop enables AFL++ to systematically explore program state space.

Instrumentation Modes

Mode Use Case Performance
afl-clang-fast (LTO) Source available, best performance Highest
afl-clang-fast Source available, standard High
afl-gcc-fast GCC-based projects High
QEMU mode Binary-only, no source Medium
Frida mode Binary-only, cross-platform Medium
Unicorn mode Firmware, embedded Low

Persistent Mode

Persistent mode avoids fork overhead by fuzzing within a loop:

c
#include <unistd.h>

__AFL_FUZZ_INIT();

int main() {
    __AFL_INIT();
    unsigned char *buf = __AFL_FUZZ_TESTCASE_BUF;

    while (__AFL_LOOP(10000)) {
        int len = __AFL_FUZZ_TESTCASE_LEN;
        // Process buf[0..len-1]
        parse_input(buf, len);
    }
    return 0;
}

Workflow

Step 1 --- Build the Fuzzing Harness

Create a harness that feeds AFL++ input to the target function:

c
// fuzz_harness.c
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "target_parser.h"

__AFL_FUZZ_INIT();

int main() {
    __AFL_INIT();
    unsigned char *buf = __AFL_FUZZ_TESTCASE_BUF;

    while (__AFL_LOOP(10000)) {
        int len = __AFL_FUZZ_TESTCASE_LEN;
        if (len < 4) continue;

        // Reset state between iterations
        parser_context_t ctx;
        parser_init(&ctx);
        parser_process(&ctx, buf, len);
        parser_cleanup(&ctx);
    }
    return 0;
}

Step 2 --- Compile with AFL++ Instrumentation

bash
# Standard instrumentation
export CC=afl-clang-fast
export CXX=afl-clang-fast++

# Enable AddressSanitizer for better crash detection
export AFL_USE_ASAN=1

# Build the target with instrumentation
$CC -o fuzz_harness fuzz_harness.c -ltarget_parser -fsanitize=address

# Build a CmpLog binary for better coverage
$CC -o fuzz_harness_cmplog fuzz_harness.c -ltarget_parser \
  -fsanitize=address -DCMPLOG

Step 3 --- Prepare Seed Corpus

bash
mkdir -p corpus/
# Add valid input samples
cp test_inputs/* corpus/
# Minimize the corpus
afl-cmin -i corpus/ -o corpus_min/ -- ./fuzz_harness @@
# Further minimize individual inputs
mkdir -p corpus_tmin/
for f in corpus_min/*; do
    afl-tmin -i "$f" -o "corpus_tmin/$(basename $f)" -- ./fuzz_harness @@
done

Step 4 --- Configure CI/CD Integration

GitHub Actions:

yaml
name: Fuzz Testing
on:
  push:
    branches: [main]
  schedule:
    - cron: '0 2 * * *'  # Nightly fuzzing

jobs:
  fuzz:
    runs-on: ubuntu-latest
    timeout-minutes: 120
    steps:
      - uses: actions/checkout@v4

      - name: Install AFL++
        run: |
          sudo apt-get update
          sudo apt-get install -y aflplusplus

      - name: Restore corpus cache
        uses: actions/cache@v4
        with:
          path: corpus/
          key: fuzz-corpus-${{ github.sha }}
          restore-keys: fuzz-corpus-

      - name: Build fuzzing harness
        run: |
          export CC=afl-clang-fast
          export AFL_USE_ASAN=1
          make fuzz_harness

      - name: Run AFL++ fuzzing (CI mode)
        env:
          AFL_CMPLOG_ONLY_NEW: 1
          AFL_FAST_CAL: 1
          AFL_NO_STARTUP_CALIBRATION: 1
        run: |
          mkdir -p findings/
          timeout 7200 afl-fuzz \
            -S ci_fuzzer \
            -i corpus/ \
            -o findings/ \
            -t 5000 \
            -- ./fuzz_harness @@ || true

      - name: Check for crashes
        run: |
          CRASHES=$(find findings/ -path "*/crashes/*" -not -name "README.txt" | wc -l)
          echo "Found $CRASHES unique crashes"
          if [ "$CRASHES" -gt 0 ]; then
            echo "::error::AFL++ found $CRASHES crashes"
            for crash in findings/*/crashes/*; do
              [ -f "$crash" ] && echo "Crash: $crash ($(wc -c < $crash) bytes)"
            done
            exit 1
          fi

      - name: Update corpus cache
        if: always()
        run: |
          afl-cmin -i findings/ci_fuzzer/queue/ -o corpus/ -- ./fuzz_harness @@

Step 5 --- Parallel Fuzzing for Nightly Runs

bash
# Launch multiple secondary instances for better coverage
for i in $(seq 1 $(nproc)); do
    afl-fuzz -S fuzzer_$i \
      -i corpus/ \
      -o findings/ \
      -- ./fuzz_harness @@ &
done

# Wait for all fuzzers
wait

# Merge and minimize corpus
afl-cmin -i findings/*/queue/ -o corpus_merged/ -- ./fuzz_harness @@

Step 6 --- Crash Triage

bash
# Reproduce and categorize crashes
for crash in findings/*/crashes/*; do
    echo "=== Testing: $crash ==="
    timeout 5 ./fuzz_harness_asan "$crash" 2>&1 | head -20
    echo "---"
done

# Deduplicate crashes by stack trace
afl-collect findings/ crashes_deduped/ -- ./fuzz_harness @@

CI/CD Best Practices for AFL++

Setting CI Short Run Nightly Long Run
Duration 30-60 min 4-24 hours
Mode -S (secondary only) -S (no -M for CI)
AFL_CMPLOG_ONLY_NEW 1 1
AFL_FAST_CAL 1 0
AFL_NO_STARTUP_CALIBRATION 1 0
Corpus caching Required Required
Parallel instances 1-2 nproc

Monitoring Fuzzing Campaigns

bash
# View fuzzing statistics
afl-whatsup findings/

# Key metrics to track:
# - Total paths found (code coverage indicator)
# - Unique crashes / unique hangs
# - Stability percentage (should be >90%)
# - Exec speed (execs/sec)
# - Cycles done (full corpus cycles completed)

References

Expand your agent's capabilities with these related and highly-rated skills.

mukul975/Anthropic-Cybersecurity-Skills

mapping-mitre-attack-techniques

Maps observed adversary behaviors, security alerts, and detection rules to MITRE ATT&CK techniques and sub-techniques to quantify detection coverage and guide control prioritization. Use when building an ATT&CK-based coverage heatmap, tagging SIEM alerts with technique IDs, aligning security controls to adversary playbooks, or reporting threat exposure to executives. Activates for requests involving ATT&CK Navigator, Sigma rules, MITRE D3FEND, or coverage gap analysis.

4,300 470
Explore
mukul975/Anthropic-Cybersecurity-Skills

hunting-for-spearphishing-indicators

Hunt for spearphishing campaign indicators across email logs, endpoint telemetry, and network data to detect targeted email attacks.

4,300 470
Explore
mukul975/Anthropic-Cybersecurity-Skills

analyzing-malicious-url-with-urlscan

URLScan.io is a free service for scanning and analyzing suspicious URLs. It captures screenshots, DOM content, HTTP transactions, JavaScript behavior, and network connections of web pages in an isolat

4,300 470
Explore
mukul975/Anthropic-Cybersecurity-Skills

implementing-zero-standing-privilege-with-cyberark

Deploy CyberArk Secure Cloud Access to eliminate standing privileges in hybrid and multi-cloud environments using just-in-time access with time, entitlement, and approval controls.

4,300 470
Explore
mukul975/Anthropic-Cybersecurity-Skills

implementing-pam-for-database-access

Deploy privileged access management for database systems including Oracle, SQL Server, PostgreSQL, and MySQL. Covers session proxy configuration, credential vaulting, query auditing, dynamic credentia

4,300 470
Explore
mukul975/Anthropic-Cybersecurity-Skills

detecting-t1003-credential-dumping-with-edr

Detect OS credential dumping techniques targeting LSASS memory, SAM database, NTDS.dit, and cached credentials using EDR telemetry, Sysmon process access monitoring, and Windows security event correlation.

4,300 470
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results