Works with the AI Agents you already use
Vargr Hunter — Threat hunting for your cloud, servers and yo
Stop guessing if your infrastructure is under attack.
$34
Vargr Hunter — Threat hunting for your cloud, servers and yo
Example session with this skill installed
We had a spike of logins this morning. Are we under attack? We have Entra ID sign-in logs, nginx logs in Loki and CloudTrail.
- Read your context and instructions
- Compiled the vargr hunter
- Generated the document
✅ Worked · Severity: P1 · Confidence: High · Status: active
Summary
Password spraying from 91[.]240[.]118[.]20 against 48 accounts, with one success: admin signed in at 10:08 UTC. 6 minutes later the same session created a new access key in AWS. Revoke now.
Covered: Entra ID sign-ins, nginx (Loki), CloudTrail · Not covered: Kubernetes audit, DNS
Do this now (you decide; nothing was changed)
☐ Revoke all sessions for admin and reset its password + MFA (Entra) — signs the user out everywhere
☐ Deactivate access key AKIA…7Q2 created at 10:14 (IAM) — anything using it will fail
☐ Block 91[.]240[.]118[.]0/24 at the IdP / WAF — may hit shared IPs
Timeline (UTC)
| Time | Event | Detection |
|---|---|---|
| 09:52 | 1,240 requests, 96% 404, sqlmap user agent | HD-WEB-03 |
| 10:05 | 1 failure per account across 48 accounts | HD-AUTH-02 |
| 10:08 | admin signs in from the same IP, no MFA (legacy auth) | HD-AUTH-04 |
| 10:14 | CreateAccessKey for user ci-deploy by admin session | HD-CLOUD-02 |
ATT&CK: T1595.002 → T1110.003 → T1078 → T1098.001
Anomaly pass: new source (first seen today), new country (RU), python-requests user agent, failures ×43.
False-positive checks
• Not our office egress, not a known scanner, no change ticket for the new key.
Attack or bug: attack (external source, payloads, enumerates accounts, unrelated to deploys).
Keep this detection
• Alert: ≥10 distinct users failing from one IP in 30 min (KQL included).
• Block legacy authentication with Conditional Access.
Next steps
• Report: .vargr/threats/2026-10-14-password-spray-admin.md · IOCs saved to .vargr/iocs.json
vargr-hunter-threat-hunting-for-your-clo.pdf
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Stop guessing if your infrastructure is under attack. Vargr Hunter performs rapid security sweeps across authentication, web, cloud, and Kubernetes logs to identify anomalies and active threats in minutes. It converts raw logs into structured P1-P4 incidents mapped to the MITRE ATT&CK framework, providing clear containment steps without ever touching your live environment.
What it does
- Multi-source sweeps analyzes AWS CloudTrail, Azure Entra ID, GCP Audit, Kubernetes audit, and WAF/API logs.
- Threat detection identifies brute force, password spraying, injection, privilege escalation, and exfiltration using 45 pre-defined logic sets.
- Local monitoring watches host processes, network connections, autostart entries, and SSH keys via Vargr Watch for macOS, Linux, and Windows.
- Incident triage groups findings by actor and ranks them by kill-chain stage, success rate, and asset criticality.
- Remediation planning generates specific containment playbooks and code-level fix recommendations for vulnerabilities.
How it works
- Scope and discover identifies connected log sources and establishes a 7-day traffic baseline to distinguish attacks from noise.
- Execute detections runs targeted queries against your observability backend or local logs using backend-neutral detection logic.
- Analyze and triage clusters findings into incidents and performs false-positive checks against known CI/CD identities and scanners.
- Report and contain delivers a threat report with an attack timeline and checkboxes for human-approved containment actions.
Frameworks & tools
The skill integrates with AWS, Azure, GCP, and Kubernetes. It supports querying via Elasticsearch (ES|QL), Splunk (SPL), KQL, CloudWatch Insights, and LogQL.
Why this beats prompting it yourself
Generic prompts often miss the subtle signals of privilege escalation or cloud control plane abuse. This skill provides a rigorous catalog of 45 security-specific detections and a formal triage process that separates true attacks from benign system spikes.
Use cases
- Conduct a 5-minute security sweep after a suspicious spike in API errors or traffic.
- Investigate whether a specific IP address or user account has successfully bypassed authentication.
- Monitor a local development machine for unauthorized processes or modified system files.
- Draft an incident response plan and containment steps during an active security event.
Known limitations
This tool is strictly read-only and cannot execute containment actions. It requires pre-existing logs or the Vargr Tracer adapter skills for most cloud and web detections.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Security.
Read-only operation: no automated containment. Requires Python 3.9+. Network log analysis depends on Vargr Tracer adapters or manual log exports. Baseline learning period is 24 hours.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 26 days ago
- Software
Python
Git