ioc-enrichment-workflow
Workflow for enriching raw IOCs. Routes each IOC type to the appropriate /lookup-* skills, optionally correlates against MISP, and synthesises a single enrichment record per indicator. Use when the us
SKILL.mdskills/ioc-enrichment-workflow/SKILL.md
GitHubIOC Enrichment Workflow
This workflow defines how to enrich raw indicators of compromise by routing them to the appropriate /lookup-* skills and synthesising results.
Seeding the queue (optional): when you don't already have a batch but want fresh leads, /lookup-crowdstrike indicators --malicious --since 7d [--type … | --actor …] returns the latest high-confidence CrowdStrike IOCs (newest-first). Feed the returned indicators into the per-type routing below to enrich each. Requires CrowdStrike credentials with the Indicators read scope.
This skill invokes: /lookup-liberty91, /lookup-virustotal, /lookup-abuseipdb, /lookup-greynoise, /lookup-shodan, /lookup-otx, /lookup-censys, /lookup-urlscan, /lookup-misp, /lookup-opencti, /lookup-reversinglabs, /lookup-crowdstrike, optionally /lookup-ransomwarelive, optionally /lookup-sentinel (exposure sweep of the confirmed-malicious subset against your own telemetry); then /score-source, /apply-tlp, /confidence-language. For deeper graph traversal, hands off to /indicator-pivoting.
Enrichment routing by IOC type
IPv4/IPv6 address
| Order | Skill | What to extract |
|---|---|---|
| 0 | /lookup-liberty91 |
First-party — run before spending third-party quota. ioc-lookup <ip> for the platform's own verdict, derived confidence, effective TLP and tags; threat-events --q <ip> if it appears in reporting |
| 1 | /lookup-virustotal |
Detection ratio, community score, associated domains, last analysis results |
| 2 | /lookup-abuseipdb |
Abuse confidence score, report count, ISP, usage type, country |
| 3 | /lookup-greynoise |
Classification (benign/malicious/unknown), noise status, actor, tags |
| 4 | /lookup-shodan |
Open ports, banners, services, OS, hostnames, vulns, last update |
| 5 | /lookup-otx |
Pulse count, associated pulses, reputation, related indicators |
| 6 | /lookup-reversinglabs |
RL classification, files seen contacting the IP (malware-corpus reputation). Run when configured — independent of VT/AbuseIPDB. |
| 7 | /lookup-crowdstrike |
indicator <ip> — Falcon Intel malicious confidence, linked actors + malware families, report refs. Run when configured. |
| 8 | /lookup-censys |
Services, certificates, autonomous system, location |
| 9 | /lookup-misp |
Internal correlation — search-attributes --value <ip> to surface prior catalogued events |
| 10 | /lookup-opencti |
Internal correlation — lookup <ip> to surface existing observables/indicators in your knowledge base |
Domain
| Order | Skill | What to extract |
|---|---|---|
| 0 | /lookup-liberty91 |
First-party — run first. ioc-lookup <domain> for verdict, derived confidence, effective TLP and tags |
| 1 | /lookup-virustotal |
Detection ratio, WHOIS, DNS records, subdomains, communicating files |
| 2 | /lookup-urlscan |
Existing scans (search, don't re-submit by default): screenshot, page content, redirects, technologies, IPs resolved |
| 3 | /lookup-shodan |
DNS resolution, open ports on resolved IPs |
| 4 | /lookup-otx |
Pulse count, associated indicators, passive DNS |
| 5 | /lookup-reversinglabs |
RL classification, files seen resolving / contacting the domain (malware-corpus reputation). Run when configured. |
| 6 | /lookup-crowdstrike |
indicator <domain> — Falcon Intel malicious confidence, linked actors + malware families, report refs. Run when configured. |
| 7 | /lookup-censys |
Certificate history, associated IPs (paid plan) |
| 8 | /lookup-misp |
search-attributes --value <domain> for internal correlation |
| 9 | /lookup-opencti |
lookup <domain> — existing observables/indicators in your knowledge base |
| 10 | /lookup-ransomwarelive |
search --q <org-candidate> — sweep ransomware leak-site claims that match the apex (see /domain-investigation § Ransomware-claim hits for caveats) |
URL
| Order | Skill | What to extract |
|---|---|---|
| 0 | /lookup-liberty91 |
First-party — run first. ioc-lookup <url> for verdict, derived confidence, effective TLP and tags |
| 1 | /lookup-virustotal |
Detection ratio, final URL, redirections, downloaded files |
| 2 | /lookup-urlscan |
Existing scans first; submit fresh only if no recent capture exists. Screenshot, DOM, requests, IPs contacted, technologies |
| 3 | /lookup-otx |
Pulse associations, reputation |
| 4 | /lookup-reversinglabs |
RL classification, files seen requesting the URL (malware-corpus reputation). Run when configured. Use submit-url only for fresh crawl + sandbox. |
| 5 | /lookup-crowdstrike |
indicator <url> — Falcon Intel malicious confidence, linked actors + malware families, report refs. Run when configured. |
| 6 | /lookup-misp |
search-attributes --value <url> for internal correlation |
| 7 | /lookup-opencti |
lookup <url> — existing observables/indicators in your knowledge base |
File hash (MD5, SHA-1, SHA-256)
| Order | Skill | What to extract |
|---|---|---|
| 0 | /lookup-liberty91 |
First-party — run first. ioc-lookup <hash> for verdict, derived confidence, effective TLP; library malware --q <family> once VT/RL name the family |
| 1 | /lookup-virustotal |
Detection ratio, file type, size, names, behavioural analysis, MITRE ATT&CK tags |
| 2 | /lookup-reversinglabs |
Run when configured — strongest single-source verdict. hash --av-scanners --ticloud for classification + threat name + AV ratio; report --detailed for MITRE ATT&CK mapping, sandbox, and networkthreatintelligence (C2 indicators extracted from the sample) |
| 3 | /lookup-crowdstrike |
indicator <hash> — Falcon Intel verdict + the threat actors and malware families CrowdStrike links to this hash + report refs. Run when configured. |
| 4 | /lookup-otx |
Pulse associations, related indicators, YARA matches |
| 5 | /lookup-misp |
search-attributes --value <hash> for internal correlation |
| 6 | /lookup-opencti |
lookup <hash> — existing observables/indicators in your knowledge base |
| 7 | /lookup-ransomwarelive |
iocs <group> and yara <group> if the hash hits a known ransomware family from VT/RL classification |
Email address
| Order | Skill | What to extract |
|---|---|---|
| 0 | /lookup-liberty91 |
First-party — run first. ioc-lookup <email> for verdict, derived confidence, effective TLP |
| 1 | /lookup-virustotal |
Associated domains and files (premium feature, may return empty on free tier) |
| 2 | /lookup-otx |
Pulse associations |
| 3 | /lookup-misp |
search-attributes --type email --value <email> for internal correlation |
| 4 | /lookup-opencti |
lookup <email> — existing observables/indicators in your knowledge base |
Enrichment process
Step 1: Parse and classify
Read the input IOC list. For each indicator:
- Determine type (IPv4, IPv6, domain, URL, hash, email)
- Validate format (regex check)
- Deduplicate
Step 2: Batch and route
Group IOCs by type. For each group, dispatch the relevant /lookup-* skills (see routing tables above).
Parallelisation: For a single IOC, dispatch all relevant lookups in parallel. For bulk IOCs, process in batches of 10 to respect rate limits (see tools/REGISTRY.md for per-API limits).
Step 3: Synthesise results
For each IOC, combine results from all lookups into a single enrichment record:
indicator: 203.0.113.42
type: ipv4-addr
enrichment_date: 2026-04-04
source_assessment: F6 # Automated enrichment, no human judgment yet
virustotal:
detection_ratio: 12/87
community_score: -45
associated_domains: [evil.example.com, bad.example.org]
abuseipdb:
abuse_confidence: 95
total_reports: 234
isp: "Bulletproof Hosting Inc"
country: RU
greynoise:
classification: malicious
noise: false
tags: [c2, cobalt-strike]
shodan:
ports: [80, 443, 8443]
os: Linux
vulns: [CVE-2024-12345]
otx:
pulse_count: 7
tags: [apt28, fancy-bear, c2]
censys:
services: [HTTP, HTTPS]
certificate_issuer: "Let's Encrypt"
misp:
matched_attributes: 3
matched_events: [42, 137] # event IDs in the local instance
prior_tags: ["tlp:amber", "actor:apt28"]
opencti:
known_as: [observable, indicator]
indicator_score: 85 # x_opencti_score on the existing indicator
prior_labels: [apt28, c2]
synthesis:
verdict: malicious
confidence: 85
context: "Known C2 server associated with APT28 campaigns. Hosted on bulletproof infrastructure in Russia. Multiple community reports plus prior MISP events confirm malicious activity."
tags: [apt28, c2, cobalt-strike]
mitre_attack: [T1071.001]
Step 4: Assess and tag
Apply source assessment (Admiralty Scale) to the enrichment with /score-source:
- Source reliability: B (established tool APIs, usually reliable)
- Information credibility: based on corroboration across tools (3+ tools agree → 1/Confirmed; 2 agree → 2/Probably true; single source → 3/Possibly true)
- A MISP or OpenCTI hit on a previously-curated event/entity lifts credibility one step (your team has already vetted it once)
- A Liberty91 hit brings its own Admiralty rating — use the occurrence's
credibilityband and each source'sreliabilitygrade verbatim instead of re-deriving them from tool agreement
Step 5: Store
Write enrichment results to data/iocs/active/YYYY-MM-DD-<context>.md with appropriate frontmatter. Apply /apply-tlp before sharing outside the team.
Step 6 (optional): Push back
If the enrichment confirms a previously-unknown malicious indicator, push it back into your own platform so future enrichments hit your catalogue first: /lookup-misp add-attribute (or create-event for a fresh cluster) for a MISP instance, /lookup-opencti create-indicator (with --score and --labels) for an OpenCTI knowledge base, and/or /lookup-liberty91 ingest to file the finding as a report in Liberty91 (it is enriched and matched into a Threat Event for your account). Liberty91 writes are metered and publish to your account — confirm with the user first.
Step 6b (optional): Sweep your own telemetry
If $SENTINEL_WORKSPACE_ID + app credentials are set, take the confirmed-malicious subset (not the whole raw list) and chain /lookup-sentinel for a batch exposure sweep — one let-bound dynamic list per IOC type, run only against tables that skill confirms exist in the workspace. This answers the question enrichment cannot: did any of these touch us? Hits become incident leads (route to the matching /*-investigation); a clean sweep is reported as "not observed in collected telemetry over <window>", never as "not compromised".
Step 7 (optional): Pivot
If the enrichment surfaces strong cluster candidates (cert siblings, JARM matches, communicating files), hand off to /indicator-pivoting for the next hop.
Rate-limit awareness
Per-API limits live in tools/REGISTRY.md. Summary:
| Service | Free-tier limit | Mitigation |
|---|---|---|
| VirusTotal (free) | 4/min, 500/day | Batch with 15s delays |
| URLScan.io (free) | 100 scans/day | Prefer search over submit |
| Shodan (free) | 1 req/sec | Sequential processing |
| AbuseIPDB (free) | 1000 checks/day | IPs only |
| GreyNoise (free) | 50 req/day | IPs only, prioritise |
| OTX | 10k req/hour | Batch freely |
| Censys | 250/month | Selective use only |
| MISP | host-bound | Local; no public limit |
| OpenCTI | host-bound | Local; no public limit |
| Liberty91 | per-key X-RateLimit-* + monthly credits |
1 credit per lookup; watch _meta.credits_remaining, and note a 429 may mean credits, not rate |
| ransomware.live (PRO) | 3000/day | Plenty for bulk org-candidate sweeps |
| ReversingLabs (A1000) | undocumented; 429 + Retry-After |
Back off on 429. Each --pivot entry on the ip op is a separate call — fan out deliberately. |
Handling missing API keys
If an API key is not configured for a service:
- Skip that enrichment source
- Note in the synthesis that the source was unavailable
- Adjust confidence accordingly (fewer sources = lower corroboration)
- Continue with available sources
To configure missing keys, point the user at /cti-setup.
Related skills
/lookup-liberty91,/lookup-virustotal,/lookup-abuseipdb,/lookup-greynoise,/lookup-shodan,/lookup-otx,/lookup-censys,/lookup-urlscan,/lookup-misp,/lookup-opencti,/lookup-reversinglabs,/lookup-crowdstrike,/lookup-ransomwarelive— the underlying lookups/lookup-sentinel— batch exposure sweep of confirmed-malicious IOCs against your own Sentinel workspace (step 6b)/ip-investigation,/domain-investigation,/hash-investigation,/url-investigation— single-seed first-hop chains; this workflow is the bulk-list equivalent/indicator-pivoting— when an enrichment opens new pivot candidates/score-source,/apply-tlp,/confidence-language— apply rigor to each finished enrichment record