QOMPASS AI

App

Redteam

Redteam

Version 1.00Source on GitHub

OVERVIEW

Redteam is Qompass AI's Red Team / Penetration Testing Educational Repository: a curated collection — not one program — of reference documents (including the NIST AI RMF Playbook), BlackArch packaging, Kali build recipes, a Flipper firmware tarball, a LaTeX pentest report, reference books, and one buildable vendored tool (Hak5's wifi-coconut), gated by a structural validation test suite.

HOW IT WORKS

  1. 1Browse the collection by area: `Gov/NIST/` (AI RMF Playbook PDF), `OGRed/` (Red Teaming Handbook, MITRE ATT&CK PDF, Hak5 tools, reference books), `Reports/` (LaTeX pentest report + audit), `BlackArch/` (packaging), `Kali/` (ARM build scripts + linux recipes), `Flipper/` (firmware tarball), `RedAI/` (AI red-teaming PDFs, primo only).
  2. 2Read the linked method references the README badges point to — Kali documentation, the OWASP Web Security Testing Guide, and pentesting tutorials — alongside the in-repo handbooks.)
  3. 3Use the helper scripts as-is: defang indicators in a file before sharing them, and generate a cvechecker config rooted in the XDG config/data/cache directories.
  4. 4Build the one buildable tool if needed: Hak5 wifi-coconut (`OGRed/Hak5/wifi-coconut`, C + CMake) configures and builds with CMake (the validation suite builds it in a scratch copy, with `-DCMAKE_POLICY_VERSION_MINIMUM=3.5` for its old `cmake_minimum_required`).
  5. 5Validate the collection after any change: `python3 -m pytest tests/ -q` checks every PDF's `%PDF`/`%%EOF` structure, the Flipper tarball's readability, the wifi-coconut build, the report's `\input` graph, and `bash -n` parsing of the BlackArch shell scripts.
How it works, step by step — each step as the project's own documentation states it.

CONCEPTS

A CURATED COLLECTION, GATED — NOT ONE PROGRAMCOLLECTION TAXONOMY — BROWSE BY AREAGov/NISTOGRedReportsBlackArchKaliFlipperRedAIGov/NIST: AI RMF Playbook · OGRed: handbook, MITRE ATT&CK, Hak5 tools, booksReports: LaTeX pentest report · BlackArch: packaging · Kali: ARM builds + recipesFlipper: firmware tarballRedAI: AI red-teaming PDFsSTRUCTURAL VALIDATION GATE — python3 -m pytest tests/ -qevery PDF: %PDF header + %%EOF markerFlipper tarball readablewifi-coconut builds (CMake, scratch copy)report \input graph · BlackArch scripts: bash -nRUN THE GATE AFTER ANY CHANGE
The collection's taxonomy — the Gov/NIST, OGRed, Reports, BlackArch, Kali, Flipper and RedAI areas — and the structural validation gate a change passes through: python3 -m pytest tests/ -q checks every PDF's structure, the Flipper tarball, the wifi-coconut build, the report's input graph, and the BlackArch shell scripts, exactly as the repo's own test suite (tests/test_collection.py) defines it.

CONTROLS

ActionControl
Validate the whole collectionpython3 -m pytest tests/ -q
Defang a file's indicators`./scripts/defang.sh input.txt` (sed: `.` → `[.]`, `@` → `[@]`, `http` → `hxxp`, `://` → `[:]//`; header shows `./sed-defang.sh input.txt > output.txt` and `-i` forms)
Generate a cvechecker configRun `scripts/cvec` — writes `$XDG_CONFIG_HOME/cvechecker/cvechecker.conf` (sqlite, XDG cache/data dirs)
Build wifi-coconut`cmake -S <src> -B <build> -DCMAKE_POLICY_VERSION_MINIMUM=3.5`, then `cmake --build <build> -j4` (as exercised by the test suite)
Install BlackArch tools (upstream doc carried in-repo)Install on an existing Arch machine, use the live ISO, or run `blackarch-install` (Full/Netinstall ISOs; Slim ISO has a GUI installer); tools install individually or in groups (2900 tools)

NOT COVERED

- No configuration surface exists to document: the repo ships documents, scripts, and vendored material with no config file, settings schema, or environment options — hence CONFIG is NONE SOURCED. The two scripts hard-code their behavior (`defang.sh` sed rules; `cvec` XDG paths + a `versions.dat` URL) rather than taking options. - The GitHub/sandbox README beyond the title is Qompass boilerplate (quick-start that clones `qompassai/dotfiles`, About/Support/FAQ); the collection description and gates trace to primo `tests/test_collection.py` and the repo tree. The sandbox clone (`~/workspace/repos/redteam`) lacks primo's `tests/`, `RedAI/`, `logs/`, and `Gov/` detail — primo is the fuller snapshot. - LaTeX report compilation is blocked by the environment, not the source (primo TeX Live lacks `pdflatex.fmt`/`texmf.cnf`, measured 2026-10-16 per the test docstring); only its `\input` graph is gated. Nothing in the repo was executed for this pack beyond reading.

REFERENCES

  1. MITRE ATT&CKMITRE's adversary tactics and techniques knowledge base; the collection carries an ATT&CK reference under OGRed/ and its red-teaming method references build on it.
  2. NIST AI Risk Management Framework PlaybookNIST's companion playbook for the AI Risk Management Framework; the collection carries the Playbook PDF under Gov/NIST/. The framework itself is at nist.gov/itl/ai-risk-management-framework.