Scenario categories
Each category is a repository of scenarios. Mappings are to MITRE ATLAS techniques and the OWASP lists; see TTP-MAP.md for the reasoning. The tools around the runner (install, judge a transcript, report, fact-check) are described in TOOLS.md. The one-command demo and the comparison of runs over time are in WORKFLOWS.md. Replaying a saved run, the defender notes and the CI check are in AUTOMATION.md.
MCP server attacks (feintkit-mcp)
Loopback ports 47100 to 47199.
| Scenario | Title | ATLAS | OWASP LLM | OWASP Agentic |
|---|---|---|---|---|
| mcp-01-tool-poisoning | Tool poisoning demo server | AML.T0110.000, AML.T0011.002 | LLM01:2025 | ASI01, ASI02 |
| mcp-02-rug-pull | Rug-pull demo server | AML.T0109, AML.T0110.001 | LLM03:2025 | ASI04 |
| mcp-03-tool-shadowing | Tool shadowing demo server | AML.T0110.000, AML.T0074 | LLM01:2025 | ASI02, ASI04 |
| mcp-04-excessive-scope | Excessive permission scope demo server | AML.T0053 | LLM06:2025 | ASI02, ASI03 |
| mcp-05-resource-injection | Indirect injection via resource content demo server | AML.T0051.001 | LLM01:2025 | ASI01 |
| mcp-06-credential-harvest | Credential harvesting via parameters demo server | AML.T0110.000, AML.T0086 | LLM02:2025 | ASI03 |
| mcp-07-command-injection-sim | Simulated command injection demo server | AML.T0050, AML.T0102 | LLM05:2025 | ASI05 |
| mcp-08-exfil-sim | Data exfiltration simulation demo server | AML.T0086, AML.T0025 | LLM02:2025 | ASI02 |
| mcp-09-impersonation | Server impersonation / typosquatting demo server | AML.T0074, AML.T0073, AML.T0010.005 | LLM03:2025 | ASI04 |
| mcp-10-sampling-abuse | MCP sampling abuse demo server | AML.T0051.001, AML.T0053 | LLM01:2025, LLM06:2025 | ASI01, ASI02 |
| mcp-11-elicitation-phish | MCP elicitation phishing demo server | AML.T0052, AML.T0073 | LLM02:2025 | ASI09 |
| mcp-12-oauth-confused-deputy | MCP OAuth confused deputy / token passthrough demo | AML.T0091.000 | none | ASI03 |
| mcp-13-exposed-local-server | Exposed local MCP server demo | AML.T0132, AML.T0006.002, AML.T0049 | none | none |
| mcp-14-cross-tool-poisoning | Cross-tool context poisoning demo (two-server) | AML.T0110.002, AML.T0080.001, AML.T0051.001 | LLM01:2025 | ASI01, ASI06 |
Agent skill attacks (feintkit-skills)
Loopback ports 47200 to 47299.
| Scenario | Title | ATLAS | OWASP LLM | OWASP Agentic |
|---|---|---|---|---|
| skill-01-prompt-injection | Skill prompt injection demo skill | AML.T0115.002, AML.T0110.000, AML.T0051.001 | LLM01:2025 | ASI01, ASI04 |
| skill-02-rug-pull | Skill rug-pull demo skill | AML.T0109, AML.T0115.002 | LLM03:2025 | ASI04 |
| skill-03-impersonation | Skill impersonation/typosquatting demo skill | AML.T0074, AML.T0073, AML.T0010.005 | LLM03:2025 | ASI04 |
| skill-04-bundled-script | Bundled malicious script demo skill | AML.T0110.001, AML.T0050, AML.T0011.002 | LLM03:2025 | ASI04, ASI05 |
| skill-05-trigger-hijack | Skill trigger hijack demo skill | AML.T0110.000 | LLM01:2025 | ASI01, ASI02 |
| skill-06-shadowing | Skill shadowing / scope precedence demo skill | AML.T0074, AML.T0110.000 | LLM03:2025 | ASI04 |
| skill-07-allowed-tools | Skill over-broad allowed-tools demo skill | AML.T0053 | LLM06:2025 | ASI03 |
| skill-08-poisoned-references | Skill poisoned reference files demo skill | AML.T0051.001, AML.T0110.000 | LLM01:2025 | ASI01 |
Prompt and context injection (feintkit-injection)
Loopback ports 47300 to 47399.
| Scenario | Title | ATLAS | OWASP LLM | OWASP Agentic |
|---|---|---|---|---|
| inj-01-web-page | Indirect injection via web page demo | AML.T0051.001, AML.T0093 | LLM01:2025 | ASI01 |
| inj-02-document | Indirect injection via email/document/PDF demo | AML.T0051.001, AML.T0093 | LLM01:2025 | ASI01 |
| inj-03-scm-text | Indirect injection via PR/issue comment demo | AML.T0051.001, AML.T0093, AML.T0119 | LLM01:2025 | ASI01 |
| inj-04-invisible-text | Invisible-text injection demo | AML.T0068, AML.T0051.001 | LLM01:2025 | ASI01 |
| inj-05-image-ocr | Image/OCR injection demo | AML.T0129, AML.T0051.001 | LLM01:2025 | ASI01 |
| inj-06-config-file | Agent config-file poisoning demo | AML.T0081, AML.T0051.001 | LLM01:2025 | ASI01, ASI06 |
| inj-07-memory | Agent memory poisoning demo | AML.T0080.000 | LLM01:2025 | ASI06 |
| inj-08-rag | RAG / vector-store poisoning demo | AML.T0070, AML.T0071, AML.T0066 | LLM08:2025, LLM04:2025 | ASI06 |
| inj-09-jailbreak-corpus | Jailbreak & system-prompt-extraction prompt corpus | AML.T0054, AML.T0056, AML.T0051.000 | LLM01:2025, LLM07:2025 | none |
| inj-12-multi-turn-split | Payload split across conversation turns | AML.T0051.000, AML.T0080.001 | LLM01:2025 | ASI01 |
| inj-13-delayed-trigger | Dormant instruction triggered several turns later | AML.T0051.001, AML.T0080.001 | LLM01:2025 | ASI01, ASI06 |
Exfiltration and output handling (feintkit-exfil)
Loopback ports 47400 to 47499.
| Scenario | Title | ATLAS | OWASP LLM | OWASP Agentic |
|---|---|---|---|---|
| exf-01-markdown-image | Markdown-image / link exfiltration demo | AML.T0077 | LLM02:2025, LLM05:2025 | none |
| exf-02-ansi-escape | ANSI/terminal escape injection demo | AML.T0067, AML.T0068 | LLM05:2025 | none |
| exf-03-insecure-output | Insecure output handling demo (dry-run sinks) | AML.T0050, AML.T0102 | LLM05:2025 | ASI05 |
Agent trust and multi-agent attacks (feintkit-agents)
Loopback ports 47500 to 47599.
| Scenario | Title | ATLAS | OWASP LLM | OWASP Agentic |
|---|---|---|---|---|
| agt-01-approval-fatigue | Approval-fatigue / UI spoofing demo | AML.T0067 | LLM06:2025 | ASI09 |
| agt-02-a2a-spoofing | Agent-to-agent message spoofing demo | AML.T0118.001, AML.T0073 | none | ASI07 |
| agt-03-computer-use-injection | Browser/computer-use agent injection demo | AML.T0100, AML.T0051.001 | LLM01:2025 | ASI01 |
| agt-04-argument-smuggling | Argument-smuggling approval demo | AML.T0053, AML.T0050 | LLM06:2025 | ASI02, ASI09 |
| agt-05-approval-bypass | Approval-rule bypass demo | AML.T0053, AML.T0050 | LLM06:2025 | ASI02, ASI05 |
| agt-06-subagent-handoff | Sub-agent hand-off demo | AML.T0118.001, AML.T0053 | LLM06:2025 | ASI07, ASI01 |
| agt-07-tool-chain-relay | Tool-chain relay demo | AML.T0053, AML.T0051.001 | LLM01:2025, LLM06:2025 | ASI02, ASI01 |
AI supply chain (feintkit-supplychain)
Loopback ports 47600 to 47699.
| Scenario | Title | ATLAS | OWASP LLM | OWASP Agentic |
|---|---|---|---|---|
| sup-01-slopsquat | Slopsquatting canary package demo | AML.T0060, AML.T0011.001, AML.T0010.001 | LLM03:2025, LLM09:2025 | ASI04 |
| sup-02-model-file | Model-file payload marker demo | AML.T0018.002, AML.T0011.000, AML.T0010.003 | LLM03:2025, LLM04:2025 | ASI04 |
| sup-03-marketplace-typosquat | Plugin/extension marketplace typosquat demo | AML.T0074, AML.T0111, AML.T0115.002 | LLM03:2025 | ASI04 |
Resource abuse and research (feintkit-abuse)
Loopback ports 47700 to 47799.
| Scenario | Title | ATLAS | OWASP LLM | OWASP Agentic |
|---|---|---|---|---|
| abu-01-denial-of-wallet | Denial-of-wallet runaway tool loop demo | AML.T0034.002, AML.T0029 | LLM10:2025 | none |
| abu-02-deferred-research | Deferred research: training-data poisoning & model extraction (decision record) | AML.T0020, AML.T0018.000, AML.T0024.002 | LLM04:2025 | none |