Autonomous Hacking Agent for Red Team
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.
Skill Claude Code
CI/CD pipeline attack category — poisoned pipeline execution, GitHub Actions expression injection, self-hosted runner abuse, secrets/OIDC exfil. Routing skill: fingerprint the CI provider + workflow surface, then load the matching leaf.
Skill Claude Code
Extracting CI secrets / OIDC tokens once you have code execution in a build job — echo/printenv exfil, log-masking bypass (base64, char-split, reversal), OIDC token abuse to assume cloud roles, GITHUBTOKEN / CIJOBTOKEN scope abuse, cache / artifact secret leakage, provenance pivot.
Skill Claude Code needs its repo
GitHub Actions ${{ }} expression injection — attacker-controlled context (issue/PR title, body, branch name, commit message) substituted into run: steps, unsafe pullrequesttarget + PR-head checkout, GITHUBTOKEN scope abuse, artifact/cache poisoning, action tag-vs-SHA pinning.
Skill Claude Code
Poisoned Pipeline Execution (PPE) — direct + indirect: inject commands via attacker-controllable build files (Makefile, package.json scripts, build.gradle, Dangerfile, .pre-commit-config.yaml), abuse pullrequesttarget / fork-PR triggers, and ride dependency / test-script execution on CI.
Skill Claude CodeCodex
Cipher detection + automated decryption via Ciphey, Cyberchef recipes, hashcat hash-ID, format conversion, common encoding chains.
Skill Claude Code
ICS / OT attack category — Modbus, BACnet, S7Comm, DNP3. Routing skill: fingerprint the industrial protocol by port + Wireshark dissector, then load the matching sub-skill. SAFETY-CRITICAL: always confirm written scope before any write/control class action.
Skill Claude Code
BACnet/IP attack — UDP/47808 discovery via Who-Is broadcast, ReadProperty / WriteProperty without auth, BBMD abuse for remote reach, vendor-specific I-Am responses, COV (Change Of Value) subscription flood, Building Automation HMI pivot.
Skill Claude Code
DNP3 attack — TCP/20000 (or 19999 serial-over-TCP) outstation enumeration, binary input / analog input poll, control relay output block (CROB) actuation, unsolicited reporting abuse, DNP3 Secure Authentication (DNP3-SA) downgrade, vendor-specific objects.
Skill Claude Code
EtherNet/IP + CIP (TCP 44818 / UDP 2222) attack playbook — List Identity broadcast, pylogix tag-database dump, tag read/write on Allen-Bradley ControlLogix/CompactLogix, CIP Forward Open, PLC mode change (Stop/Run), and historical Rockwell auth-bypass CVEs. North American ICS dominant protocol.
Skill Claude Code
Modbus TCP attack — port 502 enumeration, coil/holding-register read/write without auth, function-code abuse (FC8 diagnostic, FC43 read-device-id), Modbus-over-Serial via TCP gateway, write-with-no-confirm DoS, value tampering against PLCs.
Skill Claude Code
OPC-UA (TCP 4840) attack playbook — endpoint enumeration, SecurityPolicy mapping, anonymous/weak-auth abuse, address-space browsing and tag read, HistoryRead exfiltration, Method call for control actions, session-exhaustion DoS. Modern IT/OT DMZ convergence protocol replacing legacy fieldbus.
Skill Claude Code
PROFINET (L2 EtherType 0x8892 / DCP) attack playbook — DCP Identify-All broadcast enumeration, device fingerprinting, station-name and IP reassignment (breaks IO-controller mapping), flash-LED physical location, factory-reset, RT frame injection/replay for cyclic-IO spoofing. Siemens/EU fieldbus peer of S7Comm…
Skill Claude Code
Siemens S7 PLC attack — TCP/102 ISO-TP+S7-COMM, snap7 / python-snap7 enumeration, DB/M/E/A area read+write, PLC stop/start/run, password bypass (S7-300/400 vs S7-1200/1500 differences), CVE chain (e.g., Stuxnet's legacy primitives, CVE-2019-10936).
Skill Claude Code
Exploitation finding documentation — initial access reports, exploit chain documentation, CVSS v4.0 scoring, shell/credential inventory, detection gap analysis.
Skill Claude CodeCodex
Legitimate RMM tool abuse — deploy or hijack Atera, ScreenConnect, AnyDesk, TeamViewer for persistence, lateral movement, and C2. Leverages trusted software to evade EDR and blend with IT admin traffic.
Skill Claude Code
Supply-chain attack category — dependency confusion, typosquatting, package-registry abuse, build-pipeline poisoning, SBOM manipulation.
Skill Claude CodeCodex
Dependency confusion — publish a higher-version internal package name on public registry (npm/PyPI/Maven/Crates) to coerce CI/CD into pulling attacker code.
Skill Claude Code needs its repo
Web application exploitation — the primary category skill for all web-based attacks. This is a routing skill: read this first to identify the attack type, then load the appropriate specialized sub-skill for detailed procedures. Covers 11 technique areas across injection, file access, authentication, and API…
Skill Claude CodeCodex
Account Takeover decision tree — 9 canonical ATO paths, chaining patterns (IDOR→ATO, XSS→ATO, OAuth→ATO), MFA bypass entry points.
Skill Claude Code
Broken Function Level Authorization (BFLA) — exploit action-level access control failures where lower-privileged principals invoke admin/staff functions across REST, GraphQL, gRPC, WebSocket, and background job paths.
Skill Claude CodeCodex
Blind SQL injection under hostile WAF — manual bypass playbook for when sqlmap fails because common tokens (SUBSTRING, IF, AND, WHERE, single quotes) are filtered. Covers token-fingerprinting probe loops, arithmetic-multiplication boolean evaluation, hex-encoded literals, and exponential-probe binary search. Loaded on…
Skill Claude Code
Business logic / authentication bypass / privilege escalation — POST body field tampering (role/isadmin/usertype), 2FA bypass via response manipulation, predictable TOTP seeds, hidden authorization headers, multi-step workflow tampering. For challenges tagged businesslogic, privilegeescalation, 2fabypass, or…
Skill Claude CodeCodex
Web cache deception — trick CDN/proxy into caching authenticated responses under unauthenticated URLs, exposing PII to any visitor.
Skill Claude Code
UI redressing — missing X-Frame-Options / frame-ancestors, frame-buster bypass, drag-and-drop, cursorjacking, double-clickjacking, and sensitive-action framing.
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: