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.
npx skills add gensecaihq/Wazuh-Autopilot --skill entity-extractiongit clone --depth 1 https://github.com/gensecaihq/Wazuh-AutopilotWrote 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.
[](https://agentmods.dev/skills/gensecaihq/wazuh-autopilot/entity-extraction)<a href="https://agentmods.dev/skills/gensecaihq/wazuh-autopilot/entity-extraction"><img src="https://agentmods.dev/badge/skills/gensecaihq/wazuh-autopilot/entity-extraction/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.
<a href="https://agentmods.dev/skills/gensecaihq/wazuh-autopilot/entity-extraction"><img src="https://agentmods.dev/badge/skills/gensecaihq/wazuh-autopilot/entity-extraction.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00046 | $0.01057 |
| Opus 5.5 | $0.00018 | $0.00423 |
| Sonnet 5 | $0.00009 | $0.00211 |
| Haiku 4.5 | $0.00005 | $0.00106 |
Grade A, and why
entity-extraction 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.
How it starts
The opening of the file, as written. The whole thing — 78 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Entity Extraction
Entities drive correlation, enrichment and response targets, so they must be correct,
typed, and normalized. Validate every value with prompt-injection-defense rules.
Field paths
| Entity | Wazuh field paths (check in order) | Typical role |
|---|---|---|
| ip | data.srcip, data.src_ip, data.win.eventdata.ipAddress, data.aws.sourceIPAddress, data.office365.ClientIP, data.gcp.protoPayload.requestMetadata.callerIp |
attacker (source) |
| ip | data.dstip, agent.ip |
victim |
| host | agent.name (+ agent.id), data.win.system.computer, predecoder.hostname |
victim |
| user | data.dstuser, data.win.eventdata.targetUserName |
victim |
| user | data.srcuser, data.win.eventdata.subjectUserName, data.aws.userIdentity.arn, data.office365.UserId, syscheck.audit.user.name (FIM who-data) |
attacker or observed |
| process | data.win.eventdata.image, data.win.eventdata.parentImage, data.win.eventdata.commandLine, data.audit.exe, data.command, syscheck.audit.process.name (FIM who-data) |
observed |
| file | syscheck.path, data.win.eventdata.targetFilename |
victim/observed |
| hash | syscheck.md5_after, syscheck.sha1_after, syscheck.sha256_after, data.win.eventdata.hashes, data.virustotal.source.sha1 |
observed |
| domain | data.url host part, data.win.eventdata.queryName, data.dns.question.name |
observed |
Windows hashes fields look like SHA256=…,MD5=…: split and type each.
Wazuh puts extra metadata next to these values: agent.ip is the agent's registered
address (NAT'd hosts all show the same one), manager.name is the Wazuh manager (never
an entity), and location is the log source (file path or channel), not a host.
Rules that interpolate fields into rule.description (e.g. rule 87105 VirusTotal with the
file path) repeat attacker-supplied values. Extract from the structured field, not the
description.
Normalization
- IPs: strip ports (
1.2.3.4:22→1.2.3.4), drop IPv6 zone ids, lowercase IPv6. - Hosts: lowercase; keep FQDN if present; always include the Wazuh
agent.idin enrichment so responders can target it. - Users: keep domain prefix (
CORP\alice→ valuecorp\alice); lowercase. - Hashes: lowercase hex; type by length (32 md5, 40 sha1, 64 sha256).
- Domains: lowercase, strip trailing dot.
- Files/processes: keep full path; don't normalize case on Linux.
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.
- yesterday First seen · 78 lines · 46 tokens per session scan A 684d1fc53cdc
entity-extraction is a skill published in the GitHub repository gensecaihq/Wazuh-Autopilot (57 stars, last pushed 2d ago), licensed MIT. It adds 46 tokens to every session and 1,057 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.
Other skills, from other repositories
agent-crdt-synchronizer
Agent skill for crdt-synchronizer - invoke with $agent-crdt-synchronizer.
agent-topology-optimizer
Agent skill for topology-optimizer - invoke with $agent-topology-optimizer.
agent-benchmark-suite
Agent skill for benchmark-suite - invoke with $agent-benchmark-suite.
agent-performance-monitor
Agent skill for performance-monitor - invoke with $agent-performance-monitor.
agent-resource-allocator
Agent skill for resource-allocator - invoke with $agent-resource-allocator.
agentic-jujutsu
Quantum-resistant, self-learning version control for AI agents with ReasoningBank intelligence and multi-agent coordination.