blitzstrike - Blitz Strike — a universal MCP security-audit tool

项目说明

<di="center">

<a href=" src="./.github/assets/banner.webp" alt="Blitz Strike" width="100%" /></a>

![GitHub Release]()
![GitHub Stars]()
![TypeScript]()
![Bun]()
![MCP]()

</div>

Blitz Strike

Reconnaissance at speed. Analysis in depth. before report.

Blitz Strike is a structured penetration-testing methodology — reconnaissance,
source analysis, and — delivered as a uni MCP server. It
enumerates the attack surface (BLITZ), traces source-to-sink reachability
(EAGLE-EYE), and each finding li it is reported (STRIKE).
One server, every agent: scope enforcement to submission-ready findings in a
single run_engagement call, with the relevant exploit-tool manual attached to
every result.

> A scan hit is a hypothesis. A live test is the .
>
> Blitz Strike exists to eliminate the two most common failure modes in
> automated security assessment: false positives from surface-le
> matching, and un findings reported without li.

---

What it does

Blitz Strike is a Model Context Protocol (MCP) server (TypeScript / Bun) that
packages a 3-tier security-audit methodology as callable tools — and runs the
whole engagement ser, so a single run_engagement call works from
Claude Code, Cursor, Hermes, OpenCode, Claude Desktop, Gemini, or any MCP client.

The three tiers

Blitz Strike maps a structured penetration-testing methodology — reconnaissance,
source analysis, and — into three tool tiers executed ser.

| Tier | Name | Phase | What it does |
|---|---|---|---|
| 1 | BLITZ | Reconnaissance & attack-surface mapping | Enumerates the exposed attack surface at scale: unauthenticated entry points, dangerous sinks, and authentication boundaries. |
| 2 | EAGLE-EYE | Static analysis & data-flow tracing | Traces source-to-sink reachability and enriches findings against the escalation-chain graph. Confirms a sink is *reachable*, *unauthenticated*, and *exploitable* — not merely present. |
| 3 | STRIKE | & exploitation | Performs li (marker reflection + negati), scope enforcement, and orchestration so a finding is confirmed before it is e. |

Reconnaissance → analysis → . Nothing is reported until STRIKE
confirms it.

Beyond the three tiers, Blitz Strike also ships Web3 audit (deterministic
Solidity + executable Foundry proof), a li prober (sibling
enumeration + differential IDOR), hunting intel (CWE/C + DNS
sinkhole detection), and out-of-band proof — so IDOR/auth logic bugs, blind
injection, and smart-contract findings are all caught and proven, not just
pattern-matched.

---

Autonomous, LLM-driven

Blitz Strike is driven by the LLM — Claude, Hermes, OpenCode, Codex, or any
MCP client. The LLM is the brain (plans, routes, delegates, judges); **Blitz
Strike is the deterministic hands + knowledge + guardrails**.

A full engagement is one call, or a granular agent-orchestrated cycle:

The LLM classifies the target automatically (URL → li, filesystem
path → source pipeline), then dri → analyze → → review →
report — guided by the bundled doctrine (instructions + skills + per-step
next_steps) and fanned out across the platform's nati.

→ Autonomy & doctrine — how the LLM is steered.

Why TypeScript / Bun

Single static binary via bun build --compile — ship one executable per platform.

Zero-install distribution via bunx blitzstrike / npx blitzstrike.

MCP TypeScript SDK first-class (/sdk).

One toolchain for dev + test + build + compile.

---

Quickstart (30 seconds)

npx blitzstrike install detects every installed agent CLI (Claude Code,
Cursor, OpenCode, Codex, Hermes, Gemini, Windsurf, Copilot, Cline) and writes
the correct MCP config to each one in its nati. Restart your agent and
call run_engagement.

From source:

---

Client Configuration (works in any MCP client)

Claude Code / Desktop: claude_desktop_config.json or .mcp.json

Cursor: .cursor/mcp.json

OpenCode: .mcp.json

Hermes: mcp_servers: in config.yaml

Gemini / Copilot: native MCP config

Run blitzstrike install to print the exact snippet.

---

CLI

What doctor checks

| Check | Status you'll see |
|---|---|
| Runtime (bun/node) | OK / FAIL + fix |
| Security tools catalog | 63/130 installed, 67 on-demand |
| FOFA credentials | OK / WARN + fix |
| Data layers (chains + tools-catalog) | present / missing |

Each issue carries a fix: line — no guessing.

What install does

blitzstrike install detects which MCP client config files already exist
(Claude ~/.claude.json, Cursor ~/.cursor/mcp.json, project .mcp.json) and
merges the Blitz Strike ser in — it ne your existing
MCP servers. With no client detected, it prints the snippet for manual paste.

---

Tools

BLITZ — attack-surface triage

| Tool | Purpose |
|---|---|
| blitz_scan(path, max_files) | Scan a source tree: enumerate unauth entry points + dangerous sinks with file:line refs. |
| blitz_file(path) | Same scan, single file. |

EAGLE-EYE — deep trace

| Tool | Purpose |
|---|---|
| eagle_eye(path, symbol) | Return a function's full body, sinks in scope, and auth gates in scope. |
| eagle_grep(path, sink, max_hits) | Precision sink grep — report a hit ONLY inside a function body, flagged guarded/un-guarded. |
| enrich_scan(path, max_files) | Scan + match detected sinks to escalation chains (chains.json). |

STRIKE — + recon + orchestrate

| Tool | Purpose |
|---|---|
| strike_(url, method, data, headers, marker, timeout) | Live HTTP with marker + negati + baseline. |
| strike_resol, ) | Attach a li to a finding and advance its lifecycle. |
| scope_check(target, scope, mode) | Enforce scope before acti (no-DoS, exclusion-aware, mode-gated). |
| run_engagement(target, scope, mode, max_files) | Full 3-tier audit in ONE call — scope gate → triage → chain enrichment → findings. |
| list_chains() | List all escalation chains in the data layer. |
| fofa_search(query, size, fields) | FOFA asset index search (needs FOFA_EMAIL + FOFA_KEY). |
| n(cve_id) | C from NVD 2.0 (no key required). |
| li(url) | Modular passi recon: fingerprint, WAF/tech/, crawler, params, subdomains, Wayback, API discovery, port scan. |

CATALOG — breadth layer (tools + skills knowledge base)

| Tool | Purpose |
|---|---|
| tool_lookup(name) | Look up a tool's command + flags + install + check. |
| list_tools() | List all catalog tools, grouped by category. |
| skill_lookup(topic) | Search the skills/ playbook knowledge base by topic. |
| list_skills() | List all skill playbooks. |
| read_skill(name) | Read the full content of a playbook. |
| ensure_tool(name) | Check if a tool is installed; if not, auto-install it. |

MANUALS — deep tool reference + playbooks (wired into flow)

| Tool | Purpose |
|---|---|
| read_tool_manual(name) | Read a full deep manual for a tool (317+ manuals). |
| list_manuals() | List all manuals + playbooks. |
| read_playbook(name) | Read an engagement playbook (web-app, api-security, AD, etc.). |
| list_playbooks() | List all 17 engagement playbooks. |

PHASE 4 — , measurement, reporting

| Tool | Purpose |
|---|---|
| c(AV, AC, PR, UI, S, C, I, A) | Compute a deterministic CVSS v3.1 base score + + se (self-computed, not read from NVD). |
| dedup_findings(findings) | Collapse findings that share a root cause (sink × source × CWE) into one group per root cause. |
| generate_report(findings, format) | Emit a reproducible markdown/JSON report with summary + SHA-256 integrity hash. |
| run_enterprise_benchmark() | Run the labelled enterprise corpus and report detection rate, false-positive rate, and precision. |
| co() | Enumerate language × sink-class co (5 languages) + co ratio. |

EAGLE-EYE — taint + data-flow

| Tool | Purpose |
|---|---|
| taint_file(path) | Inter-procedural taint analysis (PHP) — reachable, sanitized, authorized sinks. |
| taint_scan(code, language) | Uni taint analysis (PHP/JS/TS/Python/Java) with source→sink tracing. |
| taint_tree(path) | Taint propagation tree — how a source flows to a sink. |
| trace_data_flow(path) | Window-based data-flow trace with sanitizer + auth-gate awareness. |
| (path) | Group reachable sinks into families. |
| list_languages() | List supported analysis languages + extensions. |

INTELLIGENCE — data layer (WAF + correlations + fuzzer)

| Tool | Purpose |
|---|---|
| detect_waf(headers, body) | Detect a WAF from response headers/body (139 signatures). |
| tech_correlation(tech) | Correlate tech to known vulns + CVEs (89 technologies). |
| c(cve) | Correlate CVE to product + targets + se (53 CVEs). |
| port_correlation(port) | Correlate port to service + attack (103 ports). |
| fuzzer_payloads(category) | Fuzzing payloads + patterns + chain rules. |
| intel_summary() | Counts of every intelligence dataset. |
| payload_lookup(topic) | Find exploit payloads (66 categories from PayloadsAllTheThings). |
| read_payload(category) | Read a full payload collection. |
| template_lookup(topic) | Find nuclei detection templates (11.9k YAML signatures). |
| list_attack_() | List the full web attack- taxonomy (34 categories, 588 ). |
| attack_(category) | Enumerate the list for one category (e.g. ssrf, business logic, ai/llm). |
| check_update() | Query the npm registry for the latest blitzstrike . |

The intelligence layer (WAF signatures, tech/CVE/port correlations, fuzzer
data, vuln ontology, exploit payloads, nuclei detection templates) is sourced
from airecon (MIT), PayloadsAllTheThings (MIT), and nuclei-templates (MIT) —
loaded at runtime and wired into the tool surface above. The attack-
taxonomy (intelligence/attack_.json) is first-party.

317 indexed tool manuals + 17 playbooks from kali-pentest (Apache-2.0) — 334
manual files, 100% of the 130-tool catalog documented. These are NOT
decoration — they are wired into the flow:

tool_lookup(name) auto-attaches the tool's full manual.

run_engagement() attaches the relevant manual per matched chain's tools field.

WEB3 — smart-contract audit + proof

| Tool | Purpose |
|---|---|
| web3_audit(code, path) | Deterministic Solidity audit — 13 bug classes (reentrancy incl. ERC721/ERC1155 callbacks, unchecked return/arithmetic, tx.origin auth, signature replay, selfdestruct, delegatecall, missing access control, timestamp dependence, spot-price oracle, unbounded loop, encodePacked collision, missing address(0) check) with CWE + se + line refs. |
| foundry_poc(class, contract) | Generate an executable Foundry test that PROVES a Web3 finding. |
| foundry_run(contract_code, poc_code, class_) | Actually run the Foundry PoC (forge test) — a PASSING test = exploit reproduced (deterministic: forge's own ). |

HUNTING — doctrine + li prober + intel

| Tool | Purpose |
|---|---|
| hunting_doctrine() | High-ROI heuristics — sibling rule, A→B signal, two-account IDOR, PoC escalation, follow-the-money + the always-rejected kill-list. |
| kill_list(type) | Deterministic always-rejected check — missing headers, self-XSS, open-redirect-alone, SSRF-DNS-only, GraphQL-introspection-alone are N/A standalone. |
| sibling_scan(base_url, endpoint, token_a, token_b) | Deterministic LI prober — sibling enumeration + differential IDOR + auth matrix + mass assignment + rate limit (closes the static-analysis gap on IDOR/auth logic bugs). |
| auto_poc(type, base_url, endpoint, param) | Turn a finding into a copy-paste-ready reproduction recipe (marker + negati + expected differential + Patchstack-style title). |
| watchlist() | Ranked CWE frequency + high-value C with payout + saturated flag — "what to hunt next". |
| resol(host) | Detect an ISP/country DNS sinkhole (blocking page) before auditing a sensiti domain. |
| oob_start() / oob_poll(id) / oob_stop(id) | Out-of-band proof — interactsh listener for blind SSRF/XXE/SQLi/command-injection (the canary must actually arrive). |

MEMORY — long-term knowledge (self-growing)

| Tool | Purpose |
|---|---|
| remember(topic, content, type, tags, ) | Save a reusable insight (deduped). =true only if marker reflected + negati inert. |
| memory_lookup(query) | Search memory by topic/tag/content, scored. |
| memory_list() | List all memory entries, grouped by type. |
| memory_forget(id) | Remove a memory entry (append-only tombstone). |
| confidence_weights(...) | Get or set the confidence-scoring weight factors. |
| finding_attach_e(finding, type, description, content) | Attach a redacted + hashed e record to a finding. |

Memory is append-only JSONL at ~/.blitzstrike/memory.jsonl (o with
BLITZSTRIKE_HOME). run_engagement auto-captures matched escalation chains
as pattern entries (deduped by chain id), so the knowledge base grows with
every engagement — no duplicate spam, and only =true entries are
authoritative.

---

Tools Catalog (tools-catalog.json)

130 self-written security tools (not copied from any project), each with command
base, key flags, install command per platform, check_installed probe, phase,
tags, alternatives, requires_root, pipes, and homepage.

| Category | Count | Examples |
|---|---|---|
| recon | 23 | subfinder, amass, httpx, naabu, katana, trufflehog |
| exploitation | 13 | sqlmap, commix, dalfox, hydra, hashcat, phpggc |
| blue-team (defensive) | 13 | suricata, zeek, osquery, wazuh, sigma, yara, trivy |
| re | 12 | ghidra, radare2, gdb, pwndbg, angr, binwalk |
| enumeration | 11 | nuclei, ffuf, gobuster, arjun, wafw00f |
| forensics | 10 | , autopsy, tshark, foremost, steghide |
| acti | 8 | netexec, impacket, bloodhound, certipy, kerbrute |
| mobile | 7 | frida, objection, mobsf, apktool, jadx |
| post-exploitation | 6 | linpeas, pspy, chisel, ligolo-ng, pwncat |
| red-team | 6 | sliver, havoc, metasploit, e, gophish |
| utility | 6 | curl, jq, anew, notify |
| web | 4 | wpscan, joomscan, droopescan, cmseek |
| wireless | 4 | aircrack-ng, wifite, bettercap |
| cloud | 4 | pacu, prowler, scoutsuite, cloudfox |
| crypto | 3 | hashid, ciphey, rsactftool |

Skills Knowledge Base (skills/)

32 skill playbooks (markdown) — uni, license-safe hidden gems from the
internet (MIT). Our personal attack-tree playbooks were removed (they were
private methodology, not uni exploit scripts).

ad (14, MIT) — ap-* prefix. Rare offensi:

cross-forest-trust-abuse, gitea-ci-injection, kerberos-trust-abuse,
multi-domain-ad-attacks, client-side-crypto-forgery.

hack.proof (18, MIT) — hp-* prefix. End-to-end audit playbooks:

full-security-audit, smart-contract-audit, api-security-test, sast-code-review,
container-image-scan, iac-cloud-posture.

See ATTRIBUTION.md for full license/copyright notices.

---

Escalation Chains (chains.json)

58 data-dri chains, each with ordered steps carrying:

tool_hint — which Blitz Strike tool to use

success_criteria — binary observable for the step

in — the assumption that MUST hold for exploitation

negati — how to refute the finding

Examples: ssrf_cloud_metadata, lfi_log_poison_rce, appkey_leak_deserialization_rce,
hmac_empty_key_forgery, int, extract_,
split_controller_upload_bypass, race_condition_double_spend, ssti_template_injection_rce,
jwt_alg_confusion_forgery, cache_poisoning_xss, subdomain_takeover, and more.

Edit chains.json to add knowledge — ne in source.

---

En

| | Required | Purpose |
|---|---|---|
| FOFA_EMAIL | For fofa_search | FOFA account email |
| FOFA_KEY | For fofa_search | FOFA API key |
| H1_USERNAME | Optional | Your HackerOne username. When set, every outbound security-testing request carries X-HackerOne-Research: <username> so targets/triagers can identify you (responsible-disclosure con). |
| BLITZSTRIKE_HOME | Optional | O the home directory (default ~/.blitzstrike). |
| BLITZSTRIKE_DATA | Optional | O the data-cache directory (default ~/.blitzstrike/data). |

All other tools need no credentials.

---

Build

---

Quality (measured, not claimed)

Blitz Strike carries a benchmark suite and a co matrix so quality is a
number you can re-run, not a claim.

Benchmark (run_enterprise_benchmark) — a labelled corpus of +

safe fixtures across PHP, Ja, Python, and Ja result:
detection rate 1.0, false-positive rate 0, precision 1.0.

Co matrix (co) — 5 languages × sink classes = 84%

co (php, ja, python, java, rust). Uncovered pairs are
legitimately absent from the language (e.g. Java has no eval, Python/Java no
PHP-style file inclusion, JS no deserialization sink).

Attack- taxonomy (list_attack_ / attack_) — a

master reference of 34 attack- categories (588 ) so a driving
agent can map the *entire* attack surface, not just the obvious sinks.

Regression suite — bun run test runs 453 checks co finding

lifecycle, e integrity, CVSS math, taint tracing, dedup, report
integrity, and benchmark in. CI runs it on every push.

---

Methodology Notes

Sink ≠ vuln. A dangerous function in the same *file* as an unauth handler does not mean the handler calls it. Use eagle_eye to confirm scope.

File write ≠ RCE (CF-003). A writable file must be *loaded by the runtime* to be execution.

Default ser only. Default Apache FilesMatch .+\.ph(ar|p|tml)$ has a $ anchor — .php.jpg does not execute.

Every finding is a HYPOTHESIS until strike_ reflects your marker AND the chain's negati stays inert.

---

Disclaimer

This tool is provided for educational and authorized security research only.
Do not use against systems without explicit permission from the owner.

---

Further Reading

Installation — prerequisites, install, client registration.

Usage — one-call engagement, granular path, modes, memory.

Autonomy & doctrine — LLM-dri, resource map, guardrails.

Manifesto — the in that drive the methodology.

Contributing — how to add chains, tools, and data.

Changelog — history.

Roadmap — where this is headed.

Security — responsible disclosure + authorized use.

Third-Party Notices — data licenses.

项目信息

开发语言：TypeScript

Star 数：500

技术要点

该项目属于防护思路方向的开源资源，代码结构清晰，适合安全研究人员学习参考。项目围绕Blitz Strike — a universal MCP security-audit toolbelt. Reconnaissance at speed. Analysis in depth. Validation before report.展开，提供了可运行的实现，对理解相关安全机制、工具原理与攻防技术具有较高的研究价值。

从技术实现角度看，此类项目的核心在于对底层原理的深入理解与工程化落地。开发者在研究过程中可以结合官方文档与社区讨论，梳理项目的初始化流程、核心模块职责与数据流转，逐步掌握其中的技术细节与设计思路。

使用与学习建议

1. 下载源码压缩包，解压到本地环境。

2. 根据项目类型准备对应的运行环境与依赖。

3. 按照 README 中的步骤完成配置与部署。

4. 在隔离的测试环境中运行验证，避免影响生产系统。

免责声明

本资源来源于开源社区，仅供安全研究、技术学习与合法授权范围内的攻防演练使用，请勿用于任何非法或侵害他人权益的用途。使用本项目产生的一切后果由使用者自行承担。