mirror of
https://github.com/mukul975/Anthropic-Cybersecurity-Skills.git
synced 2026-07-17 13:15:16 +03:00
mukul975
8cae0648ec
Demand-driven expansion targeting the fastest-growing 2025-2026 threat and
skills categories (ISC2/WEF/CrowdStrike/Mandiant signals):
- AI Security (NEW domain, 12 skills): LLM red-teaming with garak/PyRIT,
prompt injection (direct/indirect/RAG), MCP tool-poisoning, agentic tool
invocation, guardrails, model/data poisoning, system-prompt leakage,
embedding/vector weaknesses, model extraction, continuous red-teaming
- Supply Chain Security (NEW domain, 5 skills): SBOMs, dependency confusion,
malicious-npm triage, typosquatting, SLSA/Sigstore provenance
- Hardware & Firmware Security (NEW domain, 4 skills): CHIPSEC/UEFI audit,
Secure Boot bypass, TPM measured-boot attestation, ESP bootkit hunting
- Identity (10): Entra ID/ROADtools, GraphRunner, AADInternals, ADCS/Certipy,
shadow credentials, coercion, BloodHound CE, device-code phishing, SSO abuse
- Cloud-native (8): Stratus, Pacu, CloudFox, container escape, K8s RBAC,
Falco, Trivy, kube-bench
- Offensive C2 (6): Sliver, Havoc, NetExec, DPAPI, NTLM relay ESC8, redirectors
- DFIR (6): Hayabusa, Chainsaw, KAPE, Velociraptor, EZ Tools, Plaso
- Backfill (4): OpenCTI, MISP, honeytokens, post-quantum crypto migration
Each skill follows the repo taxonomy (SKILL.md + references/{standards,api-reference}.md
+ scripts/agent.py + LICENSE), with researched real tool commands (no placeholders),
complete frontmatter, and ATT&CK/ATLAS + NIST CSF mappings. Updates README domain
table, skill count, and index.json.
65 lines
2.6 KiB
Markdown
65 lines
2.6 KiB
Markdown
# API Reference — RAG Prompt Injection Testing Tools
|
|
|
|
## NVIDIA garak CLI
|
|
|
|
Install: `python -m pip install -U garak`
|
|
|
|
| Flag | Description |
|
|
|------|-------------|
|
|
| `--model_type` / `--target_type` | Generator family: `openai`, `huggingface`, `rest`, `ollama`, `bedrock` |
|
|
| `--model_name` / `--target_name` | Specific model / model id |
|
|
| `--probes` | Comma-separated probe modules, or `all` (default) |
|
|
| `--list_probes` | Print all available probe modules |
|
|
| `--generations` | Number of generations per prompt |
|
|
| `--report_prefix` | Prefix for the JSONL/HTML report |
|
|
| `--generator_option_file` | JSON config for the `rest` generator |
|
|
|
|
Relevant probe modules: `promptinject`, `latentinjection`, `leakreplay`, `xss`, `dan`, `encoding`, `goodside`.
|
|
|
|
Example:
|
|
```bash
|
|
python -m garak --model_type openai --model_name gpt-4o-mini \
|
|
--probes promptinject,latentinjection --generations 5 --report_prefix rag_run
|
|
```
|
|
|
|
## Promptfoo Red-Team CLI
|
|
|
|
Install: `npm install -g promptfoo`
|
|
|
|
| Command | Description |
|
|
|---------|-------------|
|
|
| `promptfoo redteam init` | Scaffold a red-team config |
|
|
| `promptfoo redteam run` | Generate adversarial cases and execute |
|
|
| `promptfoo redteam report` | Open the HTML findings report |
|
|
|
|
RAG-relevant plugins (`redteam.plugins`): `indirect-prompt-injection`, `rag-document-exfiltration`, `harmful:privacy`.
|
|
Strategies (`redteam.strategies`): `jailbreak`, `prompt-injection`, `crescendo`.
|
|
|
|
## Microsoft PyRIT
|
|
|
|
Install: `pip install pyrit` (Python 3.10-3.13)
|
|
|
|
| Component | Import | Purpose |
|
|
|-----------|--------|---------|
|
|
| `initialize_pyrit_async`, `IN_MEMORY` | `pyrit.setup` | Initialize memory/DB |
|
|
| `OpenAIChatTarget` | `pyrit.prompt_target` | Target an OpenAI-compatible endpoint |
|
|
| `PromptSendingAttack` | `pyrit.executor.attack` | Single-/batch-turn prompt sending |
|
|
| `RedTeamingAttack` | `pyrit.executor.attack` | Multi-turn adversarial conversation |
|
|
| `ConsoleAttackResultPrinter` | `pyrit.executor.attack` | Print scored results |
|
|
|
|
## sentence-transformers (embedding poisoning PoC)
|
|
|
|
Install: `pip install sentence-transformers faiss-cpu numpy`
|
|
|
|
| API | Description |
|
|
|-----|-------------|
|
|
| `SentenceTransformer(model_name)` | Load embedding model |
|
|
| `model.encode(text, normalize_embeddings=True)` | Produce normalized vector |
|
|
| `numpy.dot(a, b)` | Cosine similarity of normalized vectors |
|
|
|
|
## External References
|
|
|
|
- garak CLI reference: https://github.com/NVIDIA/garak/blob/main/docs/source/cliref.rst
|
|
- Promptfoo indirect injection plugin: https://www.promptfoo.dev/docs/red-team/plugins/indirect-prompt-injection/
|
|
- PyRIT docs: https://azure.github.io/PyRIT/
|