AFL++ MCP server
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@AFL++ MCP server create a new workspace named myfuzz"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
AFL++ MCP server
Model Context Protocol (MCP) server for AFL++.
This repo includes an AFLplusplus checkout (git submodule update --init with --recursive if you need AFL++ optional mode submodules) and exposes an agent-friendly API for:
creating fuzzing workspaces,
instrumenting targets,
corpus import/minimization,
harness preflight (dry run / showmap),
starting/stopping AFL++ jobs,
polling structured status and triaging findings,
other stuff
Install
Build
npm install
npm run buildInstall in Codex CLI
Build first, then register the MCP server with Codex CLI:
codex mcp add aflpp --env AFLPP_MCP_ROOT="$PWD" -- node "$PWD/dist/index.js"Run via stdio
node dist/index.jsEnvironment variables
AFLPP_MCP_ROOT(default: current working directory)AFLPP_DIR(default:$AFLPP_MCP_ROOT/AFLplusplus) – must be insideAFLPP_MCP_ROOT
Other MCP client configs
Claude Desktop
Add to your mcpServers config (adjust paths):
{
"mcpServers": {
"aflpp": {
"command": "node",
"args": ["/home/kevinv/aflpp-mcp/dist/index.js"],
"env": {
"AFLPP_MCP_ROOT": "/home/kevinv/aflpp-mcp"
}
}
}
}Related MCP server: Pwno MCP
How to use
MCP prompts
aflpp-agent-workflow: high-level end-to-end workflow (build -> corpus -> preflight -> fuzz -> triage).aflpp-harness-workplan: harness-first workflow (usage ->LLVMFuzzerTestOneInputharness -> genesis corpus -> CMPLOG/ASAN/vanilla builds -> launch commands).
MCP resources
aflpp://config: server configuration (workspace root, limits, allowlist).aflpp://docs/quickstart: some workflow notes.aflpp://docs/fuzzing_in_depth: AFL++'sfuzzing_in_depth.mdaflpp://docs/cmplog: AFL++'sinstrumentation/README.cmplog.mdaflpp://docs/env_variables: AFL++'sdocs/env_variables.mdaflpp://workspace/{name}/tree: high-level workspace treeaflpp://job/{job_name}/latest_status: latest parsed status snapshot for a jobaflpp://campaign/{campaign_name}/latest_status: latest parsed status snapshot for a campaign
MCP tools
aflpp.list_tools: List AFL++ MCP tools and their short descriptions.
aflpp.help: Get detailed help for a tool (schema + description).
aflpp.version: Get AFL++ and server version information.
aflpp.init_workspace: Create a workspace under
workspaces/<name>with standard subdirectories for inputs, outputs, targets, logs, repros, and reports.aflpp.detect_build_system: Detect a likely build system for a project path (heuristic).
aflpp.build_instrumented: Build a target with AFL++ compiler wrappers (and optional sanitizer profiles + build-time knobs) and store the artifact under the workspace
targets/directory.aflpp.build_cmplog_variant: Build a CMPLOG-instrumented variant (AFL_LLVM_CMPLOG=1) and store the artifact under the workspace
targets/directory.aflpp.import_corpus: Import a seed corpus from a file or directory into
workspaces/<ws>/in/<corpus_name>.aflpp.list_corpus: Summarize a corpus directory (file count and total size).
aflpp.list_builtin_dictionaries: List AFL++ builtin dictionaries shipped in
AFLplusplus/dictionaries.aflpp.attach_dictionary: Attach a dictionary file to a job name (stored as a job config to be used by
aflpp.start_fuzz).aflpp.dry_run: Run a short harness validation directly against the target (not
afl-fuzz) to check input mode, stability, timeouts, and basic performance.aflpp.showmap: Run
afl-showmapfor a single testcase and return a summary of the trace.aflpp.coverage_summary: Measure corpus coverage using
afl-showmap -Con an AFL++ output directory (best-effort parsing).aflpp.analyze_testcase: Run
afl-analyzeon a testcase to identify critical input regions.aflpp.preflight_checks: Run lightweight preflight checks before starting
afl-fuzz(core_pattern, CPU scaling, corpus non-empty).aflpp.start_fuzz: Start an
afl-fuzzjob in the workspace (non-blocking; supports common afl-fuzz knobs + allowlisted env overrides).aflpp.start_fuzz_cluster: Start a multi-instance
afl-fuzzcampaign (master + secondary instances; supports per-instance overrides).aflpp.stop_fuzz: Stop a running
afl-fuzzjob by PID (SIGTERM then SIGKILL).aflpp.status: Get job status by parsing
fuzzer_statsand queue/crashes/hangs counts (with deltas since last call).aflpp.campaign_summary: Summarize a multi-instance campaign by parsing
fuzzer_statsfor each instance directory.aflpp.whatsup: Run
afl-whatsupon an AFL++ output directory.aflpp.generate_progress_plot: Generate an AFL++ progress plot for a job or campaign (wraps
afl-plot).aflpp.list_findings: List crash and hang findings with stable IDs and paths.
aflpp.repro_crash: Reproduce a finding by running the target command directly with the testcase and write a repro bundle under
repros/.aflpp.crash_report: Write a crash report for a finding (dedup signature + repro info + sanitizer frames if present).
aflpp.casr_report: Generate clustered crash reports using
casr-afl(if installed).aflpp.minimize_corpus: Minimize a corpus using
afl-cminand store it as a new corpus directory in the workspace.aflpp.minimize_testcase: Minimize a single testcase using
afl-tminand store the minimized testcase underrepros/.aflpp.suggest_fuzz_cluster_mix: Suggest a multi-core campaign mix (
instance_overrides) foraflpp.start_fuzz_cluster.
Available Tools
30 toolsaflpp.analyze_testcaseC
Run afl-analyze on a testcase to identify critical input regions.
Inputs: workspace (required), target_cmd (required), testcase_path (required), timeout_ms, +3 more.
aflpp.attach_dictionaryA
Attach a dictionary file to a job name (stored as a job config to be used by start_fuzz).
Inputs: job_name (required), dict_path (required), workspace (required).
aflpp.build_cmplog_variantA
Build a CMPLOG-instrumented variant by setting AFL_LLVM_CMPLOG=1 during compilation (LLVM mode) and copy the artifact into the workspace targets/ directory.
Inputs: build_cmd (required), workspace (required), target_name (required), project_path (required), +3 more.
aflpp.build_instrumentedC
Run a constrained build command with AFL++ compiler wrappers and copy the resulting artifact into the workspace targets/ directory.
Inputs: profile (required), build_cmd (required), workspace (required), target_name (required), +4 more.
aflpp.campaign_summaryC
Summarize a multi-instance campaign by parsing fuzzer_stats for each instance directory.
Inputs: workspace (required), campaign_name (required).
aflpp.casr_reportC
Generate clustered crash reports using casr-afl (if installed).
Inputs: workspace (required), out_dir, job_name, timeout_ms, +2 more.
aflpp.coverage_summaryC
Measure corpus coverage using afl-showmap -C on an AFL++ output directory.
Inputs: workspace (required), target_cmd (required), out_dir, job_name, +4 more.
aflpp.crash_reportC
Write a crash report for a finding (dedup signature + repro info + sanitizer frames if present).
Inputs: job_name (required), workspace (required), finding_id (required), timeout_ms.
aflpp.detect_build_systemC
Detect a likely build system for a project path (heuristic).
Inputs: project_path (required).
aflpp.dry_runA
Run a short harness validation directly against the target (not afl-fuzz): checks input mode, stability, and basic performance signals.
Inputs: workspace (required), target_cmd (required), corpus_name (required), runs, +1 more.
aflpp.generate_progress_plotC
Generate an AFL++ progress plot for a job or campaign (wraps afl-plot).
Inputs: workspace (required), job_name, timeout_ms, campaign_name.
aflpp.helpA
Get detailed help for a tool (schema + description).
Inputs: tool_name (required).
aflpp.import_corpusA
Import a seed corpus into workspaces//in/<corpus_name> from a file or directory within the workspace root.
Inputs: src_path (required), workspace (required), corpus_name (required).
aflpp.init_workspaceA
Create a workspace under workspaces/ with standard subdirectories: in,out,targets,build,logs,dicts,repros,reports.
Inputs: name (required).
aflpp.list_builtin_dictionariesA
List AFL++ builtin dictionaries shipped in the AFLplusplus/dictionaries directory.
Inputs: no named parameters.
aflpp.list_corpusA
Summarize a corpus directory (file count and total size).
Inputs: workspace (required), corpus_name (required).
aflpp.list_findingsC
List crash and hang findings with stable IDs and paths.
Inputs: job_name (required), workspace (required).
aflpp.list_toolsA
List AFL++ MCP tools and their short descriptions.
Inputs: no named parameters.
aflpp.minimize_corpusB
Minimize a corpus using afl-cmin and store it as a new corpus directory in the workspace.
Inputs: workspace (required), target_cmd (required), corpus_name (required), timeout_ms, +3 more.
aflpp.minimize_testcaseB
Minimize a single testcase using afl-tmin and store the minimized testcase under repros/.
Inputs: job_name (required), workspace (required), target_cmd (required), testcase_path (required), +3 more.
aflpp.preflight_checksA
Run lightweight preflight checks before starting afl-fuzz (core_pattern, CPU scaling, corpus non-empty).
Inputs: workspace (required), target_cmd (required), corpus_name (required).
aflpp.repro_crashA
Reproduce a finding by running the target command directly with the testcase; captures stdout/stderr and writes a repro bundle under repros/.
Inputs: job_name (required), workspace (required), finding_id (required), target_cmd (required), +1 more.
aflpp.showmapC
Run afl-showmap for a single testcase and return a summary of the trace.
Inputs: workspace (required), target_cmd (required), testcase_path (required), timeout_ms, +1 more.
aflpp.start_fuzzC
Start an afl-fuzz job in the workspace (non-blocking).
Inputs: job_name (required), workspace (required), target_cmd (required), corpus_name (required), +18 more.
aflpp.start_fuzz_clusterC
Start a multi-instance afl-fuzz campaign (master + secondary instances).
Inputs: instances (required), workspace (required), target_cmd (required), corpus_name (required), +5 more.
aflpp.statusA
Get job status by parsing fuzzer_stats and queue/crashes/hangs counts (with deltas since last call).
Inputs: job_name (required), workspace (required).
aflpp.stop_fuzzA
Stop a running afl-fuzz job by PID (SIGTERM then SIGKILL).
Inputs: job_name (required), workspace (required).
aflpp.suggest_fuzz_cluster_mixC
Suggest a multi-core campaign mix (instance_overrides) for aflpp.start_fuzz_cluster.
Inputs: instances (required), master_instance_name, secondary_instance_prefix.
aflpp.versionA
Get AFL++ and server version information.
Inputs: no named parameters.
aflpp.whatsupD
Run afl-whatsup on an AFL++ output directory.
Inputs: workspace (required), out_dir, job_name, timeout_ms, +2 more.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 30 tools
Most tools have clear, distinct purposes. Minor overlap exists between status, whatsup, and campaign_summary, which all report on fuzzing progress, and among crash-related tools (repro_crash, crash_report, casr_report), but descriptions provide enough distinction.
The majority follow a verb_noun pattern with underscores (init_workspace, start_fuzz, list_findings). A few exceptions like whatsup and showmap break the pattern, and casr_report uses an abbreviation, but the overall convention is consistent.
At 30 tools, the set exceeds the 25-tool threshold for 'too many'. While the AFL++ domain is complex, several meta tools (help, version, list_tools) and overlapping status/report functions could be consolidated, making the count feel heavy.
The tool set covers the core fuzzing lifecycle: workspace setup, building, corpus management, execution, monitoring, and crash analysis. A notable gap is the lack of a 'list running jobs' tool, but most workflows are supported end-to-end.
Maintenance
Related MCP Connectors
Manage files and folders directly from your workspace. Read and write files, list directories, cre…
Hunt zero-days by talking to binaries. 40+ tools. Hosted, OAuth + SSO, invite: hi@byteray.ai
Generate SBOMs, scan vulnerabilities, and analyze dependencies from local projects or Git repos.
AI pentesting: run scans, triage vulnerabilities, review PRs, manage schedules and assets.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables interaction with Azure AI Foundry services through a unified interface for model exploration and deployment, knowledge indexing and search, AI evaluation, and fine-tuning operations. Supports both GitHub token-based model testing and full Azure deployment workflows.3MIT
- FlicenseNot gradedqualityNot gradedmaintenanceEnables stateful binary analysis and pwn research through GDB/pwndbg debugging capabilities. Supports loading binaries, setting breakpoints, memory analysis, compilation with sanitizers, and process management for vulnerability research.-
- AlicenseNot gradedqualityCmaintenanceEnables agents to operate ARL, ScopeSentry, and XingRin attack surface management platforms through stdio tools for listing resources, testing connections, creating/managing tasks, and retrieving assets.2Apache 2.0
- AlicenseNot gradedqualityNot gradedmaintenanceEnables authorized penetration testers to run reconnaissance, web fuzzing, vulnerability scanning, brute-force/password cracking, traffic capture, privilege-escalation audits, wordlist generation, and client-report workflows over stdio, HTTP, or SSE with optional bearer/OIDC authentication.157 npmGPL 3.0