PurpleAILAB/Decepticon

Autonomous Hacking Agent for Red Team

About the project

Decepticon is an autonomous red-team agent that coordinates AI agents, security tools, sandboxes, and supporting services for authorized cybersecurity assessments. Security researchers and red teams can run it through its Docker stack, cloud service, command-line interface, or Python SDK, with the catalogue entries representing its available skills.

5.5kStars on the repository
200Mods indexed here, across every type
13d agoLast push, which is what freshness is scored on
Apache-2.0Licence, which decides whether bodies are shown

terraform-state-leak

121

PurpleAILAB/Decepticon

Skill Claude CodeCodex

Exploit exposed Terraform state files — secrets, cloud creds, RDS passwords, IAM keys, and infrastructure topology in plain JSON.

not rated 5.5k +40 13d ago A SkillSpector: warn 31 tokens original Apache-2.0

contracts-overview

122

PurpleAILAB/Decepticon

Skill Claude CodeCodex

Smart contract audit lane — Solidity/EVM pattern scanner, Slither ingestion, Foundry PoC generation, DeFi attack playbooks.

not rated 5.5k +40 13d ago A SkillSpector: pass 31 tokens original Apache-2.0

access-control

123

PurpleAILAB/Decepticon

Skill Claude CodeCodex

Missing modifiers, wrong msg.sender checks, default-public functions, missing onlyOwner / onlyRole / onlyDAO authorization.

not rated 5.5k +40 13d ago A SkillSpector: pass 26 tokens original Apache-2.0

PurpleAILAB/Decepticon

Skill Claude Code

Cross-chain bridge attack — message-validation bypass (Wormhole class), validator key compromise (Ronin / Harmony class), reentrancy on token-bridge claim, Merkle-proof forgery on optimistic bridges, signature replay across chains. Bridges have lost >$2B; understand why.

not rated 5.5k +40 13d ago A SkillSpector: pass 64 tokens original Apache-2.0

flash-loan

125

PurpleAILAB/Decepticon

Skill Claude CodeCodex

Flash-loan exploit patterns — callback reentrancy, oracle amplification, governance attacks, unauthenticated callback handlers.

not rated 5.5k +40 13d ago A SkillSpector: pass 27 tokens original Apache-2.0

PurpleAILAB/Decepticon

Skill Claude Code

DAO governance attack — flash-loan-backed vote manipulation, delegation hijack, quorum dilution, proposal-spam DoS, time-lock bypass via emergency multisig, snapshot vs. on-chain vote desync, Compound/Aave/Uniswap-style GovernorBravo abuse.

not rated 5.5k +40 13d ago A SkillSpector: pass 59 tokens original Apache-2.0

PurpleAILAB/Decepticon

Skill Claude Code

MEV sandwich attacks — front-run + back-run a victim swap on Uniswap V2/V3, Curve, Balancer; mempool monitoring via Flashbots / private RPC, slippage tolerance exploitation, JIT liquidity sandwich, multi-block MEV.

not rated 5.5k +40 13d ago A SkillSpector: pass 58 tokens original Apache-2.0

oracle-manipulation

128

PurpleAILAB/Decepticon

Skill Claude CodeCodex

Hunt single-block oracle manipulation — spot-price AMM oracles, manipulable TWAP, dependent calculations, missing staleness checks.

not rated 5.5k +40 13d ago A SkillSpector: pass 33 tokens original Apache-2.0

reentrancy

129

PurpleAILAB/Decepticon

Skill Claude CodeCodex

Hunt and validate reentrancy bugs — classic cross-function, same-function, read-only, and cross-contract variants.

not rated 5.5k +40 13d ago A SkillSpector: pass 28 tokens original Apache-2.0

signature-replay

130

PurpleAILAB/Decepticon

Skill Claude CodeCodex

Signature replay attacks — missing nonces, missing chain ID, ecrecover zero address, signature malleability, cross-chain replay.

not rated 5.5k +40 13d ago A SkillSpector: pass 31 tokens original Apache-2.0

upgradeable-proxy

131

PurpleAILAB/Decepticon

Skill Claude CodeCodex

Proxy upgrade patterns and their bugs — uninitialized implementation, storage slot collisions, selector clashes, unprotected upgrade auth.

not rated 5.5k +40 13d ago B SkillSpector: warn 28 tokens original Apache-2.0

engagement-lifecycle

132

PurpleAILAB/Decepticon

Skill Claude Code

Red team engagement lifecycle management — initiation, phase transitions, go/no-go gates, deconfliction, emergency procedures, completion.

not rated 5.5k +40 13d ago A SkillSpector: pass 30 tokens original Apache-2.0

engagement-startup

133

PurpleAILAB/Decepticon

Skill Claude Code

Mandatory first-turn startup procedure — checks for existing engagements, resume/new selection, workspace initialization.

not rated 5.5k +40 13d ago A SkillSpector: pass 23 tokens original Apache-2.0

final-report

134

PurpleAILAB/Decepticon

Skill Claude Code

Final engagement report generation — executive summary, technical report, findings aggregation, attack path narrative, detection gap matrix, remediation roadmap.

not rated 5.5k +40 13d ago A SkillSpector: warn 28 tokens original Apache-2.0

kali-mcp-bridge

135

PurpleAILAB/Decepticon

Skill Claude Code

Deploy and drive Kali Linux tools via MCP-Kali-Server — structured tool-call interface, SSH tunnel setup, prompt-injection hygiene for AI-assisted engagements.

not rated 5.5k +40 13d ago C SkillSpector: warn 38 tokens original Apache-2.0

kill-chain-analysis

136

PurpleAILAB/Decepticon

Skill Claude Code

Kill chain analysis and attack path decision-making — findings analysis, attack vector selection, target prioritization, phase transitions.

not rated 5.5k +40 13d ago A SkillSpector: pass 27 tokens original Apache-2.0

orchestration

137

PurpleAILAB/Decepticon

Skill Claude Code

Decepticon orchestrator patterns — delegation, state management, adaptive re-planning, context handoff protocols.

not rated 5.5k +40 13d ago A SkillSpector: pass 26 tokens original Apache-2.0

pentest-task-tree

138

PurpleAILAB/Decepticon

Skill Claude Code

Iterative PTT (Penetration Testing Tree) session reasoning — build, update, and traverse a live numbered task tree to drive LLM-guided pentest decisions across a full session.

not rated 5.5k +40 13d ago A SkillSpector: warn 44 tokens original Apache-2.0

dfir-overview

139

PurpleAILAB/Decepticon

Skill Claude CodeCodex

Use to close the Offensive Vaccine loop on the defender side. The Detector agent produces Sigma / YARA rules from offensive operations; this catalog validates those rules against real memory dumps, event logs, and forensic artifacts using Volatility 3, plaso, and sigma-cli. Without this catalog, detection rules are…

not rated 5.5k +40 13d ago A SkillSpector: pass 68 tokens original Apache-2.0

api-overview

140

PurpleAILAB/Decepticon

Skill Claude Code

Modern API category — gRPC, SOAP/WSDL, WebSocket, Server-Sent Events. Routing skill: identify the API protocol from the response Content-Type or wire format, then load the matching sub-skill.

not rated 5.5k +40 13d ago A SkillSpector: pass 47 tokens original Apache-2.0

api-grpc

141

PurpleAILAB/Decepticon

Skill Claude Code

Skill "api-grpc" from PurpleAILAB/Decepticon, covering grpc api attack surface, discovery, 3. if reflection disabled, grab .proto from, sourcemap of a grpc-web client app and attack patterns.

not rated 5.5k +40 13d ago B SkillSpector: warn 54 tokens original Apache-2.0

PurpleAILAB/Decepticon

Skill Claude Code

Server-Sent Events (SSE / EventSource) exploitation — origin abuse for cross-site streaming exfil, prompt-injection via SSE messages into LLM clients, retry-after token leak, fragmenting events to bypass content-type sniffers.

not rated 5.5k +40 13d ago A SkillSpector: pass 54 tokens original Apache-2.0

api-soap-wsdl

143

PurpleAILAB/Decepticon

Skill Claude Code

SOAP / WSDL exploitation — WSDL enumeration via ?wsdl, XXE in SOAP envelope, WS-Addressing replay, WS-Security UsernameToken brute, SAML token injection in WS-Trust, schema validation bypass.

not rated 5.5k +40 13d ago A SkillSpector: warn 53 tokens original Apache-2.0

api-websocket

144

PurpleAILAB/Decepticon

Skill Claude Code

WebSocket exploitation — origin-bypass (CSWSH cross-site WebSocket hijacking), missing per-message auth, message-type confusion, msg-flood DoS, ws→wss downgrade, hidden RPC routes in the WS frame layer.

not rated 5.5k +40 13d ago A SkillSpector: pass 52 tokens original Apache-2.0

At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: