PPhying中文
Curated security research, vulnerabilities, advisories and tools for practitioners.

SpecterOps Skills: Open-Source LLM Agent Skills & MCP Plugin Marketplace for Red Teaming and AppSec

github.com | tool | #ai-security | #agentic-ai | #claude-code | #active-directory | #mcp | #red-team | #reverse-engineering | #ai-agents | #bloodhound | #tooling | #adcs | #tool
🔗 Original ↗

Summary

SpecterOps has open-sourced 'skills', a curated marketplace and ecosystem of LLM agent skills, plugins, and Model Context Protocol (MCP) integrations tailored for offensive security practitioners. It provides turn-key automation workflows compatible with Codex, Claude Code, and npx skills across BloodHound Active Directory analysis, C2 frameworks (Cobalt Strike, Mythic, Outflank), SCCM and ADCS attack techniques, application security reviews, and Binary Ninja reverse engineering.

Why it matters

As AI-assisted offensive security evolves from simple prompting to multi-agent architectures, SpecterOps standardizes security craft into declarative skills and MCP interfaces, allowing operators to safely orchestrate complex red team tradecraft and reverse engineering directly through LLM agents.
Published
Collected

📦 Full-Text Content Snapshot

R2 Archive · Full Article Preserved280 lines · 24.5 KB
📄 Raw .md⬇️ Download .md

Reusable agent skills, plugins, and agent definitions for SpecterOps.

Use With Codex

Each plugin includes a Codex manifest at .codex-plugin/plugin.json.

code
codex plugin marketplace add /Users/<user>/Projects/skills
# or
codex plugin marketplace add SpecterOps/skills

Then open Codex and install from /plugins.

Codex MCP setup

Codex officially supports MCP servers through declarative mcp_servers configuration. This repository no longer ships MCP runner or first-run installer scripts. Install or clone each external MCP server yourself, then point Codex at that server with command, args, and optional env values in ~/.codex/config.toml or project .codex/config.toml.

  1. Install or refresh this marketplace in Codex:
code
codex plugin marketplace add /Users/<user>/Projects/skills
# or
codex plugin marketplace add SpecterOps/skills

Then install the relevant plugins from /plugins.

  1. Configure MCP servers directly in Codex. Example BloodHound and Ghostwriter stdio configurations:
code
[mcp_servers.bloodhound_mcp]
command = "uv"
args = ["--directory", "/path/to/bloodhound-mcp", "run", "main.py"]

[mcp_servers.bloodhound_mcp.env]
BLOODHOUND_DOMAIN = "YOUR_DOMAIN"
BLOODHOUND_TOKEN_ID = "YOUR_TOKEN_ID"
BLOODHOUND_TOKEN_KEY = "YOUR_TOKEN_KEY"
BLOODHOUND_SCHEME = "https"
BLOODHOUND_PORT = "443"

[mcp_servers.ghostwriter]
command = "uv"
args = ["--directory", "/path/to/GhostWriterMCP", "run", "python", "-m", "ghostwritermcp.server"]

[mcp_servers.ghostwriter.env]
GHOSTWRITER_URL = "https://ghostwriter.example.com/"
GHOSTWRITER_API_KEY = "YOUR_API_KEY"
GHOSTWRITER_CA_BUNDLE = "/path/to/ca-bundle.crt"
GHOSTWRITER_OPLOG_ID = "123"
GHOSTWRITER_OPERATOR = "your-callsign"
GHOSTWRITER_SOURCE_IP = "10.0.0.5"
  1. Configure Binary Ninja MCP with the command or endpoint documented by your BinjaMCP installation. For stdio servers, the Codex shape is:
code
[mcp_servers.binary_ninja_mcp]
command = "npx"
args = ["-y", "binary-ninja-mcp", "--host", "localhost", "--port", "9009"]

Restart Codex after changing MCP configuration, then confirm the tools appear under /mcp before relying on MCP-assisted skills.

Use With Claude Code

Each plugin lives under plugins/<name>/ and includes a Claude Code manifest at .claude-plugin/plugin.json.

For local development:

code
/plugin marketplace add /Users/<user>/Projects/skills
/plugin install <plugin-name>@specterops-skills

For a hosted repository:

code
/plugin marketplace add SpecterOps/skills
/plugin install <plugin-name>@specterops-skills

Use With npx skills

Use npx skills when you only want to install skill instructions. This does not install full plugin behavior such as MCP config, Claude commands, hooks, or agent definitions.

code
npx skills add SpecterOps/skills --list
npx skills add SpecterOps/skills --skill <skill-name> --agent claude-code --agent codex --global

For local testing:

code
npx skills add /Users/<user>/Projects/skills --list

Contributing

Start with CONTRIBUTING.md. It covers the supported development environment, setup and validation commands, plugin metadata scaffolding, pull request expectations, and where to ask for help.

Plugins

PluginCodexClaude CodeMCPDescription
workflows-developmentYesYes-Development scaffolding, repository hygiene, security tooling, and implementation workflows for Specter Codex.
code-review-and-qaYesYes-Code review and web application QA workflows for Specter Codex.
workflows-researchYesYes-Source-backed research and synthesis workflows for Specter Codex.
ops-reconnaissanceYesYes-Reconnaissance, OSINT, service enumeration, and exposure discovery workflows for Specter Codex.
ops-appsecYesYes-Application and code security assessment workflows for Specter Codex.
ops-sccmYesYes-Microsoft Configuration Manager reconnaissance and takeover validation workflows.
bloodhoundYesYesManualBloodHound, AzureHound, GitHound/JamfHound/OktaHound OpenGraph attack-path query workflows, SCIM bridge references, and optional BloodHound MCP packaging.
payloadsYesYes-Reusable Electron payload packaging, persistence, audit, and discovery workflows.
c2-extensionsYesYes-Beacon Object File development and reusable C2 extension workflows.
report-draftingYesYesManualFinding, report drafting, Ghostwriter MCP, and operation log workflows for security assessment deliverables.
reverse-engineeringYesYesManualReverse engineering workflows and MCP-assisted binary analysis for Binary Ninja, Ghidra, and related tooling.
report-timelineYesYes-Report timeline ingestion, consolidation, and evidence packaging workflows.
codex-observabilityYesYes-Codex activity reporting and telemetry workflows.
ops-infrastructureYesYes-Infrastructure operations, SSH, tunnel, firewall, and offensive IaC attack-surface workflows.
social-engineeringYesYes-Social engineering research and phishing pretext workflows.
ludusYesYes-Ludus cyber range configuration and management skill with full API, CLI, and deployment references
c2-mythicYesYes-Mythic C2 framework implant and C2 profile development workflows with agent message protocols, payload type definitions, and listener/profile guidance.
tradecraft-windowsYes--Windows execution, persistence, and COM proxy validation workflows.
ops-adcsPlanned--Active Directory Certificate Services assessment and attack-path validation workflows. Planned; no capability is currently packaged.
ops-mssqlPlanned--Microsoft SQL Server reconnaissance, privilege mapping, and assessment workflows. Planned; no capability is currently packaged.
tradecraft-macYes--macOS execution, persistence, and operator validation workflows.
tradecraft-linuxYes--Linux process injection, execution, persistence, and operator validation workflows.
c2-cobaltstrikeYesYes-Cobalt Strike Aggressor Script, Sleep, BOF loader, and Malleable C2 profile workflows.
go-reviewYesYes-Security review for Go packages and services covering trust boundaries, injection, filesystem, crypto, concurrency, and unsafe/cgo edges
c2-outflankc2YesYes-Outflank C2 (OC2) BOF script and event-driven bot development workflows.

Skills

SkillPluginPath
git-cleanupworkflows-developmentSKILL.md
git-mergeworkflows-developmentSKILL.md
git-preflightworkflows-developmentSKILL.md
readme-generationworkflows-developmentSKILL.md
scaffold-pythonworkflows-developmentSKILL.md
scaffold-securityworkflows-developmentSKILL.md
code-reviewcode-review-and-qaSKILL.md
cpp-core-guidelinescode-review-and-qaSKILL.md
webapp-qacode-review-and-qaSKILL.md
source-researchworkflows-researchSKILL.md
nmap-parseops-reconnaissanceSKILL.md
osint-reconops-reconnaissanceSKILL.md
shodanops-reconnaissanceSKILL.md
secret-scanops-appsecSKILL.md
security-reviewops-appsecSKILL.md
webapp-reviewops-appsecSKILL.md
sccm-reconops-sccmSKILL.md
sccm-takeover-relayops-sccmSKILL.md
sccmhunter-install-localops-sccmSKILL.md
azurehound-analysisbloodhoundSKILL.md
bloodhound-ad-analysisbloodhoundSKILL.md
bloodhound-analysisbloodhoundSKILL.md
bloodhound-opengraphbloodhoundSKILL.md
bloodhound-querybloodhoundSKILL.md
openhound-developmentbloodhoundSKILL.md
openhound-githubbloodhoundSKILL.md
openhound-jamfbloodhoundSKILL.md
openhound-oktabloodhoundSKILL.md
electron-app-auditpayloadsSKILL.md
electron-candidate-discoverypayloadsSKILL.md
electron-install-persistencepayloadsSKILL.md
electron-squirrel-repackagepayloadsSKILL.md
beacon-object-file-developmentc2-extensionsSKILL.md
c2-bof-developmentc2-extensionsSKILL.md
finding-reportreport-draftingSKILL.md
ghostwriter-mcpreport-draftingSKILL.md
ghostwriter-oplogreport-draftingSKILL.md
binary-ninja-mcp-analysisreverse-engineeringSKILL.md
ghidra-mcp-analysisreverse-engineeringSKILL.md
timeline-asciinemareport-timelineSKILL.md
timeline-cobaltstrikereport-timelineSKILL.md
timeline-consolidatorreport-timelineSKILL.md
timeline-ghostwriterreport-timelineSKILL.md
timeline-markdown-notesreport-timelineSKILL.md
timeline-mythicreport-timelineSKILL.md
timeline-pdf-notesreport-timelineSKILL.md
timeline-workflowreport-timelineSKILL.md
codex-activity-reportcodex-observabilitySKILL.md
opentelemetry-codexcodex-observabilitySKILL.md
iac-attack-surfaceops-infrastructureSKILL.md
nftables-allow-sourceops-infrastructureSKILL.md
proxychains-tunnelops-infrastructureSKILL.md
ssh-opsops-infrastructureSKILL.md
credential-harvest-landing-page-copysocial-engineeringSKILL.md
phishing-campaign-buildersocial-engineeringSKILL.md
phishing-email-htmlsocial-engineeringSKILL.md
phishing-pretextsocial-engineeringSKILL.md
pretext-brainstormersocial-engineeringSKILL.md
vishing-pretextsocial-engineeringSKILL.md
ludus-developmentludusSKILL.md
mythic-implant-developmentc2-mythicSKILL.md
mythic-profilesc2-mythicSKILL.md
mythic-translation-containersc2-mythicSKILL.md
com-proxy-triagetradecraft-windowsSKILL.md
macos-initial-accesstradecraft-macSKILL.md
linux-process-injectiontradecraft-linuxSKILL.md
cobalt-strike-aggressor-developmentc2-cobaltstrikeSKILL.md
cobalt-strike-aggressor-referencec2-cobaltstrikeSKILL.md
cobalt-strike-malleable-c2-developmentc2-cobaltstrikeSKILL.md
go-reviewgo-reviewSKILL.md
oc2-bof-script-developmentc2-outflankc2SKILL.md
oc2-bot-developmentc2-outflankc2SKILL.md

Standalone Skills

SkillPath
cwe-code-reviewSKILL.md
openssf-python-reviewSKILL.md
owasp-security-code-reviewSKILL.md

Agents

AgentPath
architectagents/architect.toml
bloodhound-analystagents/bloodhound-analyst.toml
code-revieweragents/code-reviewer.toml
domain-opsagents/domain-ops.toml
exploit-devagents/exploit-dev.toml
internal-network-reconagents/internal-network-recon.toml
ludusagents/ludus.toml
mythic-developeragents/mythic-developer.toml
osint-reconagents/osint-recon.toml
planneragents/planner.toml
poc-devagents/poc-dev.toml
qa-testeragents/qa-tester.toml
report-writeragents/report-writer.toml
researcheragents/researcher.toml
reverse-engineeragents/reverse-engineer.toml
sccm-opsagents/sccm-ops.toml
security-researcheragents/security-researcher.toml
social-engineeragents/social-engineer.toml
ssh-operatoragents/ssh-operator.toml
telemetry-analystagents/telemetry-analyst.toml
winternalsagents/winternals.toml

MCP-Aware Plugins

MCP ServerPluginConfiguration
bloodhound_mcpbloodhoundConfigure directly in Codex with uv --directory /path/to/bloodhound-mcp run main.py.
ghostwriterreport-draftingConfigure directly in Codex with uv --directory /path/to/GhostWriterMCP run python -m ghostwritermcp.server.
binary_ninja_mcpreverse-engineeringConfigure directly in Codex with npx -y binary-ninja-mcp --host localhost --port 9009 after installing fosdickio/binary_ninja_mcp in Binary Ninja.
ghidrareverse-engineeringConfigure directly in Codex with the command or endpoint documented by your Ghidra MCP server.

📦 Full-Text Markdown Snapshot archived by Phying News · Original source: github.com/SpecterOps/skills

Related coverage

tool · github.com
ipatool (DrWhax Fork): App Store CLI with Native FairPlay Decryption Pipeline & MCP Integration
Significantly streamlines iOS security auditing and reverse engineering by providing an all-in-one CLI and MCP workflow to retrieve App Store packages and decrypt FairPlay binaries for static analysis.
tool · github.com
Apple Private Cloud Compute (PCC) Security Source Code and Research Tools
Provides unprecedented source-level visibility into Apple's cloud AI infrastructure, allowing security researchers to verify cryptographic privacy guarantees and audit components eligible for Apple's PCC Security Bounty.
tool · github.com
SuperGhidra: High-Performance Python Suite for Accelerated Ghidra Reverse Engineering
Drastically accelerates automated binary analysis and provides reliable mathematical verification for neural decompilers, bridging traditional reverse engineering with AI-driven workflows.
tool · github.com
KillerLand: Weaponized BYOVD Process Killer Collection Bypassing PPL & EDR Protections
Demonstrates how untrusted kernel IOCTL interfaces in signed OEM and vendor drivers continue to undermine Windows PPL security boundaries, enabling red teams to blind modern endpoint detection.
tool · github.com
StrikeAgent AtkBrain-Flash: Autonomous AI Penetration Testing & Red Teaming Platform
Demonstrates practical autonomous agent execution in offensive cybersecurity, bridging high-level LLM reasoning with real-world penetration testing toolchains for automated initial access.
research · github.com
php2shell: Pure HTTP Full-Chain RCE and ASLR Bypass on Stock PHP 8.4 Configurations
Overcomes historical constraints where PHP heap exploitation required display_errors=On, restricted memory limits, or target application vulnerabilities, demonstrating reliable pure-HTTP RCE and ASLR defeat against standard production ini configurations.
Back