Skip to content

Back to the catalog

github.comrepo

Outrider Recon

Outrider Recon is an open-source, Claude-native multi-agent OSINT and external attack-surface recon platform for authorized engagements, covering web/API, identity/SSO, cloud, secrets, and people/breach intelligence. It deliberately stops short of exploitation: active enumeration is approval-gated and intrusive validation is a human hand-off, with deterministic Python controls enforcing scope, evidence integrity, workflow state, and finding promotion around Claude skills and optional MCP-based enrichment. Output is ranked, evidence-backed finding cards (confidence/severity/handoff recommendations) meant to feed proxy-assisted testing or pen-testing rather than raw scanner dumps.

TL;DR A governed, Claude-agent-driven OSINT/attack-surface recon tool that produces ranked, evidence-backed findings under deterministic scope and approval controls, stopping before exploitation.
Document

What it is

Outrider Recon at version 4.0.0, an unsigned release candidate (tag v4.0.0) that unifies previously separate version lines (Python 0.3.0 and plugin/content 3.1.0) into one number shared by the Python package and the Claude plugin/content bundle. Beyond the concept the summary gives, the page carries the concrete install paths, the self-check contract, the security posture of the release itself, and the exact boundary of what the web plane will and won't do.

What to do

Bootstrap from a checkout — this creates an isolated venv, links Claude skills, and runs a deterministic self-check with no API keys, targets, or network needed:

git clone https://github.com/Ap6pack/outrider-recon.git
cd outrider-recon
./bootstrap.sh            # --no-web for base install, --no-skills to skip skill linking
source .venv/bin/activate
outrider                 # starts the local browser portal

A healthy self-check ends with promoted_findings_total: 0 and schema_validity_rate: 1.0 — that is the confirmation the install works before touching a target.

Docker (Linux) binds loopback only by design, so it requires host networking:

docker build -t outrider-recon .
docker run --rm --network host outrider-recon   # then open http://127.0.0.1:8765

Claude Code plugin path (skills namespaced outrider-recon:<skill>, updated via /plugin):

/plugin marketplace add Ap6pack/outrider-recon
/plugin install outrider-recon@outrider-recon

Note the CLI and web portal install separately from the skills. A one-click curl … install.sh | bash also exists, cloning to ~/.local/share/outrider-recon and symlinking skills into ~/.claude/skills/.

Caveats

  • Unsigned. Release artifacts (wheel, sdist, plugin/content bundle) are checksum-verifiable via a single SHA256SUMS, but the project claims no cryptographic signing and no PyPI / Claude Marketplace publication. The one-click installer is curl | bash; verify against the checksums yourself.
  • The web control plane is deliberately limited: it supports guarded run creation, scope checks, workflow-state transitions, approval grant/revoke, evidence-integrity verification, and explicit human finding promotion — but does not support recon execution, skill execution, automatic finding promotion, evidence editing/deletion, result upload, report generation, exports, or arbitrary MCP servers in the browser.
  • Discovery enrichment is off unless launched with outrider --enable-mcp-enrichment.
  • Origins matter for a security tool: forked from elementalsouls/Claude-OSINT, original framework SnailSploit/offensive-checklist (v1.x). MIT licensed.

Organisations named here

  • elementalsouls/Claude-OSINT
  • SnailSploit/offensive-checklist

Technologies named here

  • Outrider Recon
  • bootstrap.sh
  • install.sh
  • Claude Code plugin marketplace
  • Outrider MCP server
  • Outrider local web control plane
  • Outrider Python CLI
  • Claude-OSINT
  • offensive-checklist
  • Bellingcat's Online Investigations Toolkit
  • IntelTechniques
  • OSINT Framework
  • ProjectDiscovery
  • Six2dez reconftw
  • SecLists
  • Assetnote wordlists
  • Burp Repeater
  • Claude
  • Claude Code
  • Model Context Protocol
  • Python
  • Docker

Techniques named here

  • External asset discovery and surface mapping
  • Web and API surface recon
  • Identity fabric recon
  • Cloud and infrastructure recon
  • Secrets and dorking
  • People and breach intelligence
  • Post-credential enumeration

Related in OSINT