Agent-Skills.md

Agent Skills: OSINT Skill

Open source intelligence gathering. USE WHEN OSINT, due diligence, background check, research person, company intel, investigate. SkillSearch('osint') for docs.

UncategorizedID: Steffen025/pai-opencode/OSINT

Repository

Steffen025/pai-opencode
Steffen025License: MIT
7811

Install this agent skill to your local

pnpm dlx add-skill https://github.com/Steffen025/pai-opencode/tree/HEAD/.opencode/skills/OSINT

Skill Files

Browse the full folder contents for OSINT.

Download Skill

Loading file tree…

.opencode/skills/OSINT/SKILL.md

Skill Metadata

Name
OSINT
Description
"Open source intelligence gathering. USE WHEN OSINT, due diligence, background check, research person, company intel, investigate. SkillSearch('osint') for docs."

Customization

Before executing, check for user customizations at: ~/.opencode/skills/CORE/USER/SKILLCUSTOMIZATIONS/OSINT/

If this directory exists, load and apply any PREFERENCES.md, configurations, or resources found there. These override default behavior. If the directory does not exist, proceed with skill defaults.

OSINT Skill

Open Source Intelligence gathering for authorized investigations.

Voice Notification

When executing a workflow, do BOTH:

  1. Send voice notification:

    curl -s -X POST http://localhost:8888/notify \
  -H "Content-Type: application/json" \
  -d '{"message": "Running the WORKFLOWNAME workflow from the OSINT skill"}' \
  > /dev/null 2>&1 &

  2. Output text notification:

    Running the **WorkflowName** workflow from the **OSINT** skill...

Workflow Routing

| Investigation Type | Workflow | Context | |-------------------|----------|---------| | People lookup | Workflows/PeopleLookup.md | PeopleTools.md | | Company lookup | Workflows/CompanyLookup.md | CompanyTools.md | | Investment due diligence | Workflows/CompanyDueDiligence.md | CompanyTools.md | | Entity/threat intel | Workflows/EntityLookup.md | EntityTools.md |

Trigger Patterns

People OSINT:

  • "do OSINT on [person]", "research [person]", "background check on [person]"
  • "who is [person]", "find info about [person]", "investigate this person" -> Route to Workflows/PeopleLookup.md

Company OSINT:

  • "do OSINT on [company]", "research [company]", "company intelligence"
  • "what can you find about [company]", "investigate [company]" -> Route to Workflows/CompanyLookup.md

Investment Due Diligence:

  • "due diligence on [company]", "vet [company]", "is [company] legitimate"
  • "assess [company]", "should we work with [company]" -> Route to Workflows/CompanyDueDiligence.md

Entity/Threat Intel:

  • "investigate [domain]", "threat intelligence on [entity]", "is this domain malicious"
  • "research this threat actor", "check [domain]", "analyze [entity]" -> Route to Workflows/EntityLookup.md

Authorization (REQUIRED)

Before ANY investigation, verify:

  • [ ] Explicit authorization from client
  • [ ] Clear scope definition
  • [ ] Legal compliance confirmed
  • [ ] Documentation in place

STOP if any checkbox is unchecked. See EthicalFramework.md for details.

Resource Index

| File | Purpose | |------|---------| | EthicalFramework.md | Authorization, legal, ethical boundaries | | Methodology.md | Collection methods, verification, reporting | | PeopleTools.md | People search, social media, public records | | CompanyTools.md | Business databases, DNS, tech profiling | | EntityTools.md | Threat intel, scanning, malware analysis |

Integration

Automatic skill invocations:

  • Research Skill - Parallel researcher agent deployment (REQUIRED)
  • Recon Skill - Technical infrastructure reconnaissance

Agent fleet patterns:

  • Quick lookup: 4-6 agents
  • Standard investigation: 8-16 agents
  • Comprehensive due diligence: 24-32 agents

Researcher types: | Researcher | Best For | |------------|----------| | PerplexityResearcher | Current web data, social media, company updates | | DeepResearcher | Academic depth, professional backgrounds | | GeminiResearcher | Multi-perspective, cross-domain connections | | GrokResearcher | Contrarian analysis, fact-checking |

File Organization

Active investigations:

~/.opencode/MEMORY/WORK/$(jq -r '.work_dir' ~/.opencode/MEMORY/STATE/current-work.json)/scratch/YYYY-MM-DD-HHMMSS_osint-[target]/

Archived reports:

~/.opencode/History/research/YYYY-MM/[target]-osint/

Ethical Guardrails

ALLOWED: Public sources only - websites, social media, public records, search engines, archived content

PROHIBITED: Private data, unauthorized access, social engineering, purchasing breached data, ToS violations

See EthicalFramework.md for complete requirements.

Version: 2.0 (Canonical Structure) Last Updated: December 2024