detection-engineering

detection-engineering is a skill for Claude Code from gensecaihq/Wazuh-Autopilot. It costs 59 tokens per session (1,295 once invoked), scanned A, original, MIT.

A procedure for designing, testing, and tuning security detections as documented rules. It uses Sigma, a portable format for describing detections, or Wazuh rules, and sends proposals for human review rather than claiming they are deployed.

In plain words
What is it for?
Use it when a hunt finds activity with no alert, an existing rule creates repeated false alarms, an incident was detected late, or new attacker behaviour needs coverage.
Why use it?
It turns detection changes into reviewed, testable work instead of undocumented rule edits. It also records expected false alarms, blind spots, data sources, and validation steps.

Skill for Claude Code

Written for Claude Code: allowed-tools in frontmatter.

Good fit Use it when a hunt finds activity with no alert, an existing rule creates repeated false alarms, an incident was detected late, or new attacker behaviour needs coverage.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/gensecaihq/wazuh-autopilot/detection-engineering
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add gensecaihq/Wazuh-Autopilot --skill detection-engineering
Clone the repo
git clone --depth 1 https://github.com/gensecaihq/Wazuh-Autopilot

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for detection-engineering

README.md
[![agentmods](https://agentmods.dev/badge/skills/gensecaihq/wazuh-autopilot/detection-engineering/github.svg)](https://agentmods.dev/skills/gensecaihq/wazuh-autopilot/detection-engineering)
Your own site
<a href="https://agentmods.dev/skills/gensecaihq/wazuh-autopilot/detection-engineering"><img src="https://agentmods.dev/badge/skills/gensecaihq/wazuh-autopilot/detection-engineering/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for detection-engineering

Your own site · 80×15
<a href="https://agentmods.dev/skills/gensecaihq/wazuh-autopilot/detection-engineering"><img src="https://agentmods.dev/badge/skills/gensecaihq/wazuh-autopilot/detection-engineering.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 59 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,295 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00059 $0.01295
Opus 5.5 $0.00024 $0.00518
Sonnet 5 $0.00012 $0.00259
Haiku 4.5 $0.00006 $0.00129

Measured yesterday against content hash 83e6b5f61a35, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-26, from the pricing page.

Security

Grade A, and why

detection-engineering scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

backend/app/skills/detection-engineering/SKILL.md · 104 lines

How it starts

The opening of the file, as written. The whole thing — 104 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Detection Engineering

Detections are code: versioned, documented, tested, reviewed. You propose; humans review and deploy. Never claim a rule is deployed.

Inputs that trigger work

  • Hunt found activity with no alert (detection gap).
  • Incident where first detection was late (dwell time > 0).
  • Rule producing repeated false positives (tuning).
  • New TTP from threat intel relevant to the environment.

ADS (Alerting and Detection Strategy) — Palantir framework

Every proposal documents:

  1. Goal — what behaviour it detects.
  2. Categorization — ATT&CK tactic/technique.
  3. Strategy Abstract — how, at a high level.
  4. Technical Context — data sources, fields, platform details.
  5. Blind Spots and Assumptions — what it misses, what must be true.
  6. False Positives — known benign triggers.
  7. Validation — how to generate a true positive to test it.
  8. Priority — alert severity and why.
  9. Response — what the analyst should do when it fires.

Rule formats

Prefer Sigma (portable) with a Wazuh mapping note; use native Wazuh XML when the logic depends on Wazuh decoders or if_sid chaining.

Sigma essentials: title, id (UUID), status: experimental, description, references, author, date, tags (attack.t1110.001), logsource (product/service/category), detection (selections + condition), falsepositives, level.

Wazuh mapping notes:

  • Custom rules live in local_rules.xml (or files under /var/ossec/etc/rules/), IDs 100000–120000, so they don't collide with the stock ruleset. Write them as custom rule 100210 in findings.
  • Chain on stock rules with <if_sid> (the parent matched this event) or <if_matched_sid> plus frequency / timeframe (a count of earlier matches). Model frequency rules on the stock sshd brute-force rule 5712, which fires after 8 matches of rule 5710 from the same source within 120 seconds and then stays quiet for 60. As a custom rule 100210 that reads:
    <rule id="100210" level="10" frequency="8" timeframe="120" ignore="60">
      <if_matched_sid>PARENT_SID</if_matched_sid>
      <same_source_ip />
      <description>...</description>
    </rule>
    
    Other correlation options in the stock ruleset: <same_user />, <same_field>, <if_matched_group>.
  • Map Sigma fields to decoded Wazuh fields (data.win.eventdata.commandLine, data.win.system.eventID, data.srcip, data.dstuser) with <field name="..."> PCRE2 or OS_Regex patterns. Field names drop the data. prefix inside rules (<field name="win.eventdata.commandLine">).
  • Add <mitre><id>T1110.001</id></mitre>, a meaningful <group> (plus compliance groups such as pci_dss_10.2.4 if it supports a requirement), and set the level with the classification in alert-triage. Level 12+ pages people.
  • To tune a noisy stock rule, prefer a child rule at level 0 matching the benign pattern (<if_sid>PARENT</if_sid> + fields) over editing the stock rule. Changes to stock files are lost on upgrade.
  • Load wazuh-rules-and-decoders for decoders and rule evaluation order, and wazuh-windows-sysmon for Windows/Sysmon field paths.

Read the full file on GitHub · 104 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 104 lines · 59 tokens per session scan A 83e6b5f61a35

Subscribe to this mod's changes

detection-engineering is a skill published in the GitHub repository gensecaihq/Wazuh-Autopilot (57 stars, last pushed 2d ago), licensed MIT. It adds 59 tokens to every session and 1,295 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-25.