AGENT SKILL

performing-threat-hunting-with-yara-rules

Use YARA pattern-matching rules to hunt for malware, suspicious files, and indicators of compromise across filesystems and memory dumps. Covers rule authoring, yara-python scanning, and integration with threat intel feeds.

License 许可
Apache-2.0
最近更新
2026-08-25
Stars
★ 0
Forks
0

01 前言 Front Matter

name
performing-threat-hunting-with-yara-rules
tags
[ "yara", "malware-detection", "threat-hunting", "pattern-matching" ]
author
mahipal
domain
cybersecurity
license
Apache-2.0
version
1.0
nist_csf
[ "DE.CM-01", "DE.AE-02", "DE.AE-07", "ID.RA-05" ]
subdomain
threat-hunting
description
Use YARA pattern-matching rules to hunt for malware, suspicious files, and indicators of compromise across filesystems and memory dumps. Covers rule authoring, yara-python scanning, and integration with threat intel feeds.
mitre_attack
[ "T1046", "T1057", "T1082", "T1083", "T1005" ]
d3fend_techniques
[ "Executable Denylisting", "Execution Isolation", "File Metadata Consistency Validation", "Content Format Conversion", "File Content Analysis" ]

02 资源文件 3

03 SKILL.md 正文

Performing Threat Hunting with YARA Rules

Scan files, directories, and memory dumps using YARA rules to identify malware families, suspicious patterns, and IOC matches.

When to Use

  • Proactively hunting for unknown malware variants across network shares, endpoints, and email attachments
  • Scanning quarantine directories or sandbox outputs for malware family classification
  • Searching process memory dumps for injected code or in-memory-only payloads
  • Validating threat intelligence IOCs against a large corpus of collected samples
  • Triaging incident response artifacts to identify known malware families quickly
  • Building automated detection pipelines that scan new files on ingestion

Do not use for real-time endpoint protection (use EDR agents instead); YARA scanning is best suited for batch hunting, triage, and post-collection analysis where scan latency is acceptable.

Prerequisites

  • YARA 4.x installed (apt install yara on Debian/Ubuntu, brew install yara on macOS)
  • Python 3.8+ with yara-python (pip install yara-python)
  • yarGen for automated rule generation (git clone https://github.com/Neo23x0/yarGen)
  • Sample malware corpus or suspicious files for scanning (from malware zoos, VT, or incident artifacts)
  • Optional: pefile for PE header analysis, malduck for memory carving
  • Threat intel YARA rule sets (e.g., YARA-Rules community repository, Florian Roth signature-base)

Workflow

Step 1: Install YARA and Python Bindings

# Linux
sudo apt update && sudo apt install -y yara

# Python bindings
pip install yara-python

# Verify installation
yara --version
python3 -c "import yara; print(yara.YARA_VERSION)"

Step 2: Write a Basic YARA Rule

Create rules that match on strings, hex patterns, and file metadata:

// File: rules/emotet_loader.yar
rule Emotet_Loader_2026 {
    meta:
        author = "Threat Intel Team"
        description = "Detects Emotet first-stage loader DLL"
        date = "2026-01-20"
        reference = "https://attack.mitre.org/software/S0367/"
        mitre_attack = "T1059.001, T1055.001"
        severity = "critical"

    strings:
        // Emotet export function name patterns
        $export1 = "DllRegisterServer" ascii
        $export2 = "RunDLL" ascii nocase

        // Obfuscated string decryption routine
        $decrypt_loop = { 8B 45 ?? 33 45 ?? 89 45 ?? 8B 4D ?? 03 4D ?? }

        // PowerShell download cradle in embedded script
        $ps_cradle = /powershell[^\n]{0,50}-e(nc|ncodedcommand)/i

        // Known C2 URI patterns
        $uri1 = "/wp-content/uploads/" ascii
        $uri2 = "/wp-admin/css/" ascii
        $uri3 = "/wp-includes/" ascii

        // PE characteristics
        $mz = "MZ" at 0

    condition:
        $mz and
        filesize < 2MB and
        (
            ($export1 and $decrypt_loop) or
            ($ps_cradle and any of ($uri*)) or
            (2 of ($uri*) and $decrypt_loop)
        )
}

Step 3: Write Advanced Rules with Modules

Use YARA modules for PE header inspection and math-based entropy checks:

import "pe"
import "math"

rule Suspicious_Packed_Executable {
    meta:
        author = "Threat Hunting Team"
        description = "Detects PE files with high entropy sections indicating packing or encryption"
        severity = "medium"

    condition:
        pe.is_pe and
        pe.number_of_sections > 0 and
        for any section in pe.sections : (
            math.entropy(section.offset, section.size) > 7.2 and
            section.size > 1024
        ) and
        pe.imports("kernel32.dll", "VirtualAlloc") and
        pe.imports("kernel32.dll", "VirtualProtect")
}

rule Suspicious_UPX_Modified {
    meta:
        description = "Detects UPX-packed binaries with tampered section names"
        severity = "medium"

    strings:
        $upx_magic = { 55 50 58 21 }  // UPX!

    condition:
        pe.is_pe and
        $upx_magic and
        not (
            pe.sections[0].name == "UPX0" and
            pe.sections[1].name == "UPX1"
        )
}

Step 4: Scan Files and Directories with yara-python

import yara
import os
import json
from datetime import datetime
from pathlib import Path

def compile_rules(rule_paths):
    """Compile YARA rules from one or more .yar files."""
    rule_files = {}
    for i, path in enumerate(rule_paths):
        namespace = Path(path).stem
        rule_files[namespace] = path
    return yara.compile(filepaths=rule_files)

def scan_directory(rules, target_dir, recursive=True):
    """Scan a directory for matches and return structured results."""
    results = []
    scan_count = 0
    error_count = 0

    for root, dirs, files in os.walk(target_dir):
        for filename in files:
            filepath = os.path.join(root, filename)
            scan_count += 1
            try:
                matches = rules.match(filepath, timeout=60)
                if matches:
                    for match in matches:
                        result = {
                            "file": filepath,
                            "rule": match.rule,
                            "namespace": match.namespace,
                            "tags": match.tags,
                            "meta": match.meta,
                            "strings": [],
                            "scan_time": datetime.utcnow().isoformat()
                        }
                        for offset, identifier, data in match.strings:
                            result["strings"].append({
                                "offset": hex(offset),
                                "identifier": identifier,
                                "data": data.hex() if isinstance(data, bytes) else data
                            })
                        results.append(result)
                        print(f"  MATCH: {match.rule} -> {filepath}")
            except yara.TimeoutError:
                error_count += 1
                print(f"  TIMEOUT scanning {filepath}")
            except yara.Error as e:
                error_count += 1

        if not recursive:
            break

    print(f"\nScan complete: {scan_count} files scanned, "
          f"{len(results)} matches, {error_count} errors")
    return results

# Compile and scan
rules = compile_rules([
    "rules/emotet_loader.yar",
    "rules/suspicious_packed.yar"
])

matches = scan_directory(rules, "/mnt/evidence/collected_samples/")

# Export results
with open("yara_scan_results.json", "w") as f:
    json.dump(matches, f, indent=2)

Step 5: Scan Process Memory Dumps

Hunt for in-memory indicators that only exist in running processes:

import yara

def scan_memory_dump(rules, dump_path):
    """Scan a process memory dump for YARA matches."""
    matches = rules.match(dump_path, timeout=120)

    for match in matches:
        print(f"Rule: {match.rule}")
        print(f"  Severity: {match.meta.get('severity', 'unknown')}")
        for offset, identifier, data in match.strings:
            # Show context around the match
            print(f"  String {identifier} at offset {hex(offset)}")
            if len(data) <= 64:
                print(f"    Data: {data.hex()}")

    return matches

# Rules targeting in-memory artifacts
memory_rules = yara.compile(source="""
rule Cobalt_Strike_Beacon_Memory {
    meta:
        description = "Detects Cobalt Strike beacon in process memory"
        severity = "critical"
    strings:
        $config_start = { 2E 2F 2E 2F 2E 2C }
        $sleep_mask = { 48 8B 44 24 ?? 48 89 44 24 ?? 48 8B 44 24 }
        $named_pipe = "\\\\\\\\.\\\\pipe\\\\msagent_" ascii
        $watermark = { 00 00 00 00 00 00 ?? ?? 00 00 }
    condition:
        2 of them
}
""")

scan_memory_dump(memory_rules, "/mnt/evidence/lsass_dump.dmp")

Step 6: Generate Rules Automatically with yarGen

Use yarGen to create rules from malware samples by extracting unique strings:

# Clone and set up yarGen
git clone https://github.com/Neo23x0/yarGen.git
cd yarGen
pip install -r requirements.txt

# Download the string databases (run once)
python3 yarGen.py --update

# Generate rules from a directory of malware samples
python3 yarGen.py \
    -m /mnt/evidence/malware_samples/ \
    -o generated_rules.yar \
    --excludegood \
    -p "AutoGen" \
    -a "Threat Hunting Team" \
    --score 50

# Generate rules for a single sample with maximum detail
python3 yarGen.py \
    -m /mnt/evidence/malware_samples/suspicious.exe \
    -o single_sample_rule.yar \
    --opcodes \
    --debug

Step 7: Integrate Community Rule Sets

Download and combine rules from public threat intelligence repositories:

# Clone Florian Roth's signature-base (large community rule set)
git clone https://github.com/Neo23x0/signature-base.git

# Clone YARA-Rules community repository
git clone https://github.com/Yara-Rules/rules.git yara-community-rules

# Clone ReversingLabs YARA rules
git clone https://github.com/reversinglabs/reversinglabs-yara-rules.git
import yara
from pathlib import Path

def load_rule_directory(rule_dir, extensions=(".yar", ".yara")):
    """Load all YARA rules from a directory tree."""
    rule_files = {}
    for ext in extensions:
        for rule_file in Path(rule_dir).rglob(f"*{ext}"):
            namespace = rule_file.stem
            # Avoid namespace collisions
            if namespace in rule_files:
                namespace = f"{rule_file.parent.name}_{namespace}"
            rule_files[namespace] = str(rule_file)

    print(f"Loading {len(rule_files)} rule files from {rule_dir}")
    try:
        compiled = yara.compile(filepaths=rule_files)
        return compiled
    except yara.SyntaxError as e:
        print(f"Syntax error in rules: {e}")
        # Fall back to loading rules one by one, skipping broken ones
        valid_rules = {}
        for ns, path in rule_files.items():
            try:
                yara.compile(filepath=path)
                valid_rules[ns] = path
            except yara.SyntaxError:
                print(f"  Skipping broken rule: {path}")
        return yara.compile(filepaths=valid_rules)

# Load and scan with community rules
community_rules = load_rule_directory("signature-base/yara/")
matches = community_rules.match("/mnt/evidence/suspicious_file.exe", timeout=120)

for m in matches:
    print(f"Matched: {m.rule} (namespace: {m.namespace})")

Step 8: Build a Continuous Hunting Pipeline

Automate scanning of new files as they arrive using filesystem monitoring:

import yara
import time
import json
import hashlib
from pathlib import Path
from watchdog.observers import Observer
from watchdog.events import FileSystemEventHandler

class YaraHuntingHandler(FileSystemEventHandler):
    def __init__(self, rules, alert_file="yara_alerts.jsonl"):
        self.rules = rules
        self.alert_file = alert_file
        self.scanned_hashes = set()

    def on_created(self, event):
        if event.is_directory:
            return
        self._scan_file(event.src_path)

    def _scan_file(self, filepath):
        # Deduplicate by file hash
        try:
            file_hash = hashlib.sha256(Path(filepath).read_bytes()).hexdigest()
        except (PermissionError, FileNotFoundError):
            return

        if file_hash in self.scanned_hashes:
            return
        self.scanned_hashes.add(file_hash)

        matches = self.rules.match(filepath, timeout=60)
        if matches:
            alert = {
                "timestamp": time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime()),
                "file": filepath,
                "sha256": file_hash,
                "matches": [
                    {"rule": m.rule, "severity": m.meta.get("severity", "unknown")}
                    for m in matches
                ]
            }
            with open(self.alert_file, "a") as f:
                f.write(json.dumps(alert) + "\n")
            print(f"ALERT: {filepath} matched {len(matches)} rules")

# Set up continuous monitoring
rules = yara.compile(filepaths={"hunting": "rules/all_hunting_rules.yar"})
handler = YaraHuntingHandler(rules)
observer = Observer()
observer.schedule(handler, path="/mnt/quarantine/", recursive=True)
observer.start()
print("YARA hunting pipeline active. Monitoring /mnt/quarantine/ ...")

Verification

  • Compile all custom rules without syntax errors: yara -w rules/*.yar /dev/null
  • Confirm rules match known-good malware samples from your test corpus (true positive validation)
  • Verify rules do NOT match a goodware corpus of common system files (false positive testing)
  • Test scanning performance: single file scan should complete within timeout threshold
  • Validate yarGen output rules compile and produce meaningful matches against the input samples
  • Check that community rule sets load without critical syntax errors after filtering
  • Confirm the continuous hunting pipeline generates alerts in JSONL format when test files are dropped
  • Cross-reference YARA matches against VirusTotal or sandbox results to validate detection accuracy

04 同仓库更多 Skills

vikasvardhanv
vikasvardhanv/jerry_skills

syllabus

Generates a curated supplementary reading list from any course syllabus using Consensus academic search. Grill-me intake (syllabus input format + course audience + year range) plus a grouping forcing-options checkpoint before any search runs — so the reading list matches the course's level and recency need. Parses the syllabus to extract topics and learning outcomes, searches Consensus for recent peer-reviewed papers per topic, and produces a professionally formatted .docx with clickable Consensus links, plain-language summaries calibrated to audience level, and Bloom-higher-order discussion questions tied to course learning goals. Use when the user uploads a syllabus, course outline, or curriculum document and wants supplementary readings (e.g., 'create a reading list from this syllabus', 'find recent papers for my course') — even casual mentions with a syllabus attached should trigger this skill.

★ 0
vikasvardhanv
vikasvardhanv/jerry_skills

pulse

Multi-source recency research skill that takes the pulse of any topic across Reddit, Hacker News, the open web, and optionally X/Twitter within a configurable recent window (default 30 days). Forcing intake clarifies topic specificity, angle (trend/sentiment/problems/opportunities/comparison), time window, and platform scope before searching. Returns a synthesized briefing with citations, engagement metrics, and cross-platform pattern analysis. Use when the user requests multi-source recency intelligence on a topic (e.g., 'pulse on [topic]', 'what's happening with [topic]', 'what are people saying about [topic]', 'current conversation about [topic]', 'take the pulse of [topic]', 'trending: [topic]', 'find me info on [topic]'), and for competitor research, trend discovery, tool comparisons, and audience sentiment analysis.

★ 0
vikasvardhanv
vikasvardhanv/jerry_skills

patent

Patent prior-art and landscape intelligence skill — not generic patent help. Commits to one of five sub-use-cases via forcing intake (novelty search / freedom-to-operate / competitive landscape / acquisition diligence / litigation prior-art) before any search runs. Searches Google Patents, Espacenet, USPTO, and optionally Lens.org for citation-graph signals. Output is an editable Word document (.docx) with verdict, ranked closest art (claim-text extracted), CPC-class-aware landscape, family-resolved hits, geographic coverage, FTO flags where applicable, strategy recommendations, and full audit log. Use when the user asks for patent searching or analysis (e.g., 'prior art search for [invention]', 'freedom to operate analysis for [product]'). Produces search signal, not legal advice — always recommends consulting a patent attorney before filing or licensing decisions. Trademark, copyright, and trade-secret questions are out of scope.

★ 0
vikasvardhanv
vikasvardhanv/jerry_skills

notebooklm

Browser automation skill for controlling Google's NotebookLM. Use when the user wants anything done in NotebookLM (e.g., 'open NotebookLM', 'check my [name] notebook', 'ask my notebook about X', 'add [source] to NotebookLM', 'generate a Video Overview from my notebook', 'use NotebookLM Studio'). Handles reading and querying notebooks, adding sources (URLs, text, files, YouTube links, synthesized content), generating Studio outputs (Audio/Video Overviews, Mind Maps, Reports incl. Briefing Doc/Study Guide/FAQ, Flashcards, Quiz, slide decks, infographics — discover the exact set from the live Studio panel; the UI evolves fast), and creating new notebooks. Requires browser automation environment — fails gracefully when unavailable.

★ 0
vikasvardhanv
vikasvardhanv/jerry_skills

litreview

Academic literature orientation skill that searches papers via free keyless APIs (PubMed E-utilities + OpenAlex) by default — with the Consensus MCP as an optional enhancement lane when connected — builds a strategic search plan using PICO (default) or SPIDER / Decomposition / hybrid as fallbacks, and synthesizes findings into a formatted Word (.docx) research guide. Grill-me intake (research question specificity + framework hint + tentative depth) before the recon search; a second forcing checkpoint after Phase 2 confirms framework + sub-areas + depth before searches consume budget. Configurable depth (5/10/20 queries) controls coverage vs. speed. Output is a 'launching pad' — an orientation guide that lets a researcher dive in confidently, not a finished review. Use when the user starts literature-oriented research (e.g., 'litreview on [topic]', 'literature review on [topic]', 'I'm starting a literature review on X', 'I'm writing a paper on X', 'help me research X', 'I'm doing research on X', 'can you help me research X'). Do NOT use for single one-off paper searches wanting a quick list — that's a plain PubMed/OpenAlex (or Consensus) query.

★ 0
vikasvardhanv
vikasvardhanv/jerry_skills

grants

NIH grant research skill for clinical researchers. Grill-me intake (research idea + career stage + preliminary data + environment + submission posture + known institute targets) locks down the funding strategy before any search runs. Runs a 5-facet Consensus positioning analysis (with draft Significance/Innovation language), maps the research to the right NIH institutes and study sections via RePORTER, finds NOSIs and funded overlap, and produces an editable Word document (.docx) with budget/scope-aware mechanism recommendations, submission timelines, and a mandatory program officer recommendation. Use when the user asks about research funding or makes any grant-related request (e.g., 'grants for [topic]', 'find grants for my research idea', 'what grants match my research', 'help me find NIH funding', 'grant opportunities for my research'). NIH-only scope — non-NIH funders (PCORI, DOD CDMRP, VA, foundations) are out of scope and flagged at intake.

★ 0