README.md
GitHuboffensive-osint skill
The "what to reach for" operational arsenal for external red-team OSINT and bug-bounty reconnaissance.
| Field | Value |
|---|---|
| Name | offensive-osint |
| Version | 2.1 |
| Lines | ~3,800 |
| Top-level sections | 51 |
| Subsections | ~135 |
| Companion skill | osint-methodology |
When this skill triggers
Auto-triggers on prompts containing any of ~110 trigger phrases. Common ones:
- All triggers from
osint-methodology(most prompts pull both) swagger discovery,openapi discovery,graphql introspection,graphql field suggestionsubdomain enumeration,subdomain takeover,cloud bucket enum,S3 enum,GCS enum,Azure blob enumokta enum,entra enum,azure AD enum,ADFS enum,SAML metadatamobile recon,APK analysis,Microsoft 365 deep,Teams federation,SharePoint enum,OneDrive enumsecret scanning,secret leak,leaked credential,JWT triage,AWS key triagegithub dorking,google dorking,postman workspace,stack exchange OSINTbreach lookup,have I been pwned,HudsonRock cavalier,infostealer,dehashed,intelxshodan recon,censys recon,certificate transparency,crt.sh,JARM,favicon mmh3JS endpoint extraction,sourcemap leakcopy paste probes,curl one-lineremail security analysis,SPF DMARC DKIMorigin discovery,CDN bypass,WAF bypassvendor product fingerprints,Citrix Netscaler,F5 BIG-IP,Pulse Secure,FortiGate,PaloAlto GlobalProtect,Cisco AnyConnect,VMware vCentercloud native fingerprint,Lambda function URL,Cloud Runkubernetes exposure,kubelet,etcdCI CD exposure,Jenkins recon,GitLab self-hosted,GitHub Actions secretsdocumentation leak,Notion public,Confluence anonymous,Trello boardWHOIS RDAP,DNS record catalog,Wayback CDXLinkedIn enumeration,job posting tech stackSlack workspace discovery,Discord server discoverynpm token leak,PyPI token leak,Docker Hub leaksat imagery physical reconTLS deep audit,JA3 JA4,reverse DNS sweep,IPv6 enumerationCVE prioritization,EPSS scoring,CISA KEV,vulnerability prioritizationtooling installsector specific recon,healthcare DICOM,finance SWIFT,ICS SCADA,Modbus,BACnetpost discovery workflowAnthropic API key,OpenAI API key
Full trigger list in the SKILL.md frontmatter.
What's in it
See the parent README's "What's in the box" table for the full §-by-§ breakdown.
Highlights:
- §16 — Pre-built wordlists & probe paths including 28 Swagger paths, 13 GraphQL paths + introspection POST body, 35 high-risk ports, 6 missing security headers, 15 always-on HTTP checks, 5 SAML metadata paths, 8 SSO subdomain prefixes, cloud-bucket arsenal (6 prefixes × 15 suffixes × 47 stems × 3 providers), JS guess-paths, endpoint-extraction regex tiers, internal-host leakage regexes, 27 takeover provider fingerprints, copy-paste curl probes, email security analysis, origin discovery / CDN bypass, vendor product fingerprints, cloud-native fingerprints, container/K8s exposure, CI/CD exposure, doc/wiki leak paths, WHOIS/RDAP, DNS catalog with TXT verification token table, Wayback CDX deep usage.
- §17 — Secret-pattern catalog (48 patterns) with severity, category, false-positive notes.
- §18 — Dork corpus (80+ templates, 9 categories).
- §20 — Endpoint interest score (0–100 rubric).
- §21 — Mobile app ownership confidence (0–100 rubric).
- §22 — Identity-fabric concrete endpoints (incl. M365 Deep + GraphQL field-suggestion enum).
- §23 — 9 read-only secret validators + post-discovery enumeration workflows.
- §39 — 27 attack-path hint templates.
- §40 — Severity decision matrix (80+ worked examples).
- §41–§47 — LinkedIn enum, job posting analysis, Slack/Discord discovery, package registry leaks, sat imagery, tooling install, sector notes.
- §48 — Runnable secret-scan helper (stdlib-only Python, available standalone at
scripts/secret_scan.py).
Loading
# Local Claude Code install
cp SKILL.md ~/.claude/skills/offensive-osint/SKILL.md
cp scripts/secret_scan.py ~/.claude/skills/offensive-osint/scripts/secret_scan.py
# Or attach to a Claude.ai project / Claude API system prompt
The full content lives in this SKILL.md (or in docs/full-skills/offensive-osint.SKILL.full.md if this file is the structured-outline variant).
Helper script
scripts/secret_scan.py — stdlib-only Python scanner mirroring the §17 secret-pattern catalog. Run standalone:
python3 scripts/secret_scan.py path/to/repo/ # scan a directory tree
python3 scripts/secret_scan.py file1 file2 file3 # scan specific files
cat my.log | python3 scripts/secret_scan.py # pipe stdin
Output: JSONL — one finding per line — jq-friendly.
Self-test
Run the prompts in ../../tests/smoke-test-prompts.md. Arsenal-targeted prompts are tagged in the test file.
License
MIT — see LICENSE.