merge with main from pyrit

build_scripts/check_links.py

@@ -104,7 +104,10 @@ def check_links_in_file(file_path):
 if __name__ == "__main__":
     files = sys.argv[1:]
     all_broken_urls = {}
+    skipped_files = ["doc/blog/"]
     for file_path in files:
+        if any(file_path.startswith(skipped) for skipped in skipped_files):
+            continue
         print(f"Checking links in {file_path}")
         broken_urls = check_links_in_file(file_path)
         if broken_urls:

doc/_config.yml

@@ -32,6 +32,7 @@ html:
   use_issues_button: true
   use_repository_button: true
   use_edit_page_button: true
+  extra_static_files: ["_static/custom.js"]
 
 sphinx:
   extra_extensions:

doc/_static/custom.js

@@ -0,0 +1,244 @@
+window.addEventListener("DOMContentLoaded", () => {
+	const headerButtons = document.querySelector(".article-header-buttons");
+
+	if (headerButtons) {
+		const discordBtn = document.createElement("a");
+
+		discordBtn.title = "Join our Discord server";
+		discordBtn.href = "https://discord.com/invite/9fMpq3tc8u";
+		discordBtn.target = "_blank";
+		discordBtn.className = "btn btn-sm";
+
+		/* The Discord SVG icon is from Bootstrap Icons: https://icons.getbootstrap.com/icons/discord/ */
+		discordBtn.innerHTML = `<span class="btn__icon-container"><svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" fill="currentColor" class="svg-inline--fa" viewBox="0 0 16 16"><path d="M13.545 2.907a13.2 13.2 0 0 0-3.257-1.011.05.05 0 0 0-.052.025c-.141.25-.297.577-.406.833a12.2 12.2 0 0 0-3.658 0 8 8 0 0 0-.412-.833.05.05 0 0 0-.052-.025c-1.125.194-2.22.534-3.257 1.011a.04.04 0 0 0-.021.018C.356 6.024-.213 9.047.066 12.032q.003.022.021.037a13.3 13.3 0 0 0 3.995 2.02.05.05 0 0 0 .056-.019q.463-.63.818-1.329a.05.05 0 0 0-.01-.059l-.018-.011a9 9 0 0 1-1.248-.595.05.05 0 0 1-.02-.066l.015-.019q.127-.095.248-.195a.05.05 0 0 1 .051-.007c2.619 1.196 5.454 1.196 8.041 0a.05.05 0 0 1 .053.007q.121.1.248.195a.05.05 0 0 1-.004.085 8 8 0 0 1-1.249.594.05.05 0 0 0-.03.03.05.05 0 0 0 .003.041c.24.465.515.909.817 1.329a.05.05 0 0 0 .056.019 13.2 13.2 0 0 0 4.001-2.02.05.05 0 0 0 .021-.037c.334-3.451-.559-6.449-2.366-9.106a.03.03 0 0 0-.02-.019m-8.198 7.307c-.789 0-1.438-.724-1.438-1.612s.637-1.613 1.438-1.613c.807 0 1.45.73 1.438 1.613 0 .888-.637 1.612-1.438 1.612m5.316 0c-.788 0-1.438-.724-1.438-1.612s.637-1.613 1.438-1.613c.807 0 1.451.73 1.438 1.613 0 .888-.631 1.612-1.438 1.612" /></svg></span>`;
+
+		headerButtons.prepend(discordBtn);
+	}
+});
+(function(cfg) {
+	function e() {
+        cfg.onInit && cfg.onInit(n)
+    }
+    var x, w, D, t, E, n, C = window, O = document, b = C.location, q = "script", I = "ingestionendpoint", L = "disableExceptionTracking", j = "ai.device.";
+    "instrumentationKey"[x = "toLowerCase"](),
+    w = "crossOrigin",
+    D = "POST",
+    t = "appInsightsSDK",
+    E = cfg.name || "appInsights",
+    (cfg.name || C[t]) && (C[t] = E),
+    n = C[E] || function(g) {
+        var f = !1
+          , m = !1
+          , h = {
+            initialize: !0,
+            queue: [],
+            sv: "8",
+            version: 2,
+            config: g
+        };
+        function v(e, t) {
+            var n = {}
+              , i = "Browser";
+            function a(e) {
+                e = "" + e;
+                return 1 === e.length ? "0" + e : e
+            }
+            return n[j + "id"] = i[x](),
+            n[j + "type"] = i,
+            n["ai.operation.name"] = b && b.pathname || "_unknown_",
+            n["ai.internal.sdkVersion"] = "javascript:snippet_" + (h.sv || h.version),
+            {
+                time: (i = new Date).getUTCFullYear() + "-" + a(1 + i.getUTCMonth()) + "-" + a(i.getUTCDate()) + "T" + a(i.getUTCHours()) + ":" + a(i.getUTCMinutes()) + ":" + a(i.getUTCSeconds()) + "." + (i.getUTCMilliseconds() / 1e3).toFixed(3).slice(2, 5) + "Z",
+                iKey: e,
+                name: "Microsoft.ApplicationInsights." + e.replace(/-/g, "") + "." + t,
+                sampleRate: 100,
+                tags: n,
+                data: {
+                    baseData: {
+                        ver: 2
+                    }
+                },
+                ver: undefined,
+                seq: "1",
+                aiDataContract: undefined
+            }
+        }
+        var n, i, t, a, y = -1, T = 0, S = ["js.monitor.azure.com", "js.cdn.applicationinsights.io", "js.cdn.monitor.azure.com", "js0.cdn.applicationinsights.io", "js0.cdn.monitor.azure.com", "js2.cdn.applicationinsights.io", "js2.cdn.monitor.azure.com", "az416426.vo.msecnd.net"], o = g.url || cfg.src, r = function() {
+            return s(o, null)
+        };
+        function s(d, t) {
+            if ((n = navigator) && (~(n = (n.userAgent || "").toLowerCase()).indexOf("msie") || ~n.indexOf("trident/")) && ~d.indexOf("ai.3") && (d = d.replace(/(\/)(ai\.3\.)([^\d]*)$/, function(e, t, n) {
+                return t + "ai.2" + n
+            })),
+            !1 !== cfg.cr)
+                for (var e = 0; e < S.length; e++)
+                    if (0 < d.indexOf(S[e])) {
+                        y = e;
+                        break
+                    }
+            var n, i = function(e) {
+                var a, t, n, i, o, r, s, c, u, l;
+                h.queue = [],
+                m || (0 <= y && T + 1 < S.length ? (a = (y + T + 1) % S.length,
+                p(d.replace(/^(.*\/\/)([\w\.]*)(\/.*)$/, function(e, t, n, i) {
+                    return t + S[a] + i
+                })),
+                T += 1) : (f = m = !0,
+                s = d,
+                !0 !== cfg.dle && (c = (t = function() {
+                    var e, t = {}, n = g.connectionString;
+                    if (n)
+                        for (var i = n.split(";"), a = 0; a < i.length; a++) {
+                            var o = i[a].split("=");
+                            2 === o.length && (t[o[0][x]()] = o[1])
+                        }
+                    return t[I] || (e = (n = t.endpointsuffix) ? t.location : null,
+                    t[I] = "https://" + (e ? e + "." : "") + "dc." + (n || "services.visualstudio.com")),
+                    t
+                }()).instrumentationkey || g.instrumentationKey || "",
+                t = (t = (t = t[I]) && "/" === t.slice(-1) ? t.slice(0, -1) : t) ? t + "/v2/track" : g.endpointUrl,
+                t = g.userOverrideEndpointUrl || t,
+                (n = []).push((i = "SDK LOAD Failure: Failed to load Application Insights SDK script (See stack for details)",
+                o = s,
+                u = t,
+                (l = (r = v(c, "Exception")).data).baseType = "ExceptionData",
+                l.baseData.exceptions = [{
+                    typeName: "SDKLoadFailed",
+                    message: i.replace(/\./g, "-"),
+                    hasFullStack: !1,
+                    stack: i + "\nSnippet failed to load [" + o + "] -- Telemetry is disabled\nHelp Link: https://go.microsoft.com/fwlink/?linkid=2128109\nHost: " + (b && b.pathname || "_unknown_") + "\nEndpoint: " + u,
+                    parsedStack: []
+                }],
+                r)),
+                n.push((l = s,
+                i = t,
+                (u = (o = v(c, "Message")).data).baseType = "MessageData",
+                (r = u.baseData).message = 'AI (Internal): 99 message:"' + ("SDK LOAD Failure: Failed to load Application Insights SDK script (See stack for details) (" + l + ")").replace(/\"/g, "") + '"',
+                r.properties = {
+                    endpoint: i
+                },
+                o)),
+                s = n,
+                c = t,
+                JSON && ((u = C.fetch) && !cfg.useXhr ? u(c, {
+                    method: D,
+                    body: JSON.stringify(s),
+                    mode: "cors"
+                }) : XMLHttpRequest && ((l = new XMLHttpRequest).open(D, c),
+                l.setRequestHeader("Content-type", "application/json"),
+                l.send(JSON.stringify(s)))))))
+            }, a = function(e, t) {
+                m || setTimeout(function() {
+                    !t && h.core || i()
+                }, 500),
+                f = !1
+            }, p = function(e) {
+                var n = O.createElement(q)
+                  , e = (n.src = e,
+                t && (n.integrity = t),
+                n.setAttribute("data-ai-name", E),
+                cfg[w]);
+                return !e && "" !== e || "undefined" == n[w] || (n[w] = e),
+                n.onload = a,
+                n.onerror = i,
+                n.onreadystatechange = function(e, t) {
+                    "loaded" !== n.readyState && "complete" !== n.readyState || a(0, t)
+                }
+                ,
+                cfg.ld && cfg.ld < 0 ? O.getElementsByTagName("head")[0].appendChild(n) : setTimeout(function() {
+                    O.getElementsByTagName(q)[0].parentNode.appendChild(n)
+                }, cfg.ld || 0),
+                n
+            };
+            p(d)
+        }
+        cfg.sri && (n = o.match(/^((http[s]?:\/\/.*\/)\w+(\.\d+){1,5})\.(([\w]+\.){0,2}js)$/)) && 6 === n.length ? (d = "".concat(n[1], ".integrity.json"),
+        i = "@".concat(n[4]),
+        l = window.fetch,
+        t = function(e) {
+            if (!e.ext || !e.ext[i] || !e.ext[i].file)
+                throw Error("Error Loading JSON response");
+            var t = e.ext[i].integrity || null;
+            s(o = n[2] + e.ext[i].file, t)
+        }
+        ,
+        l && !cfg.useXhr ? l(d, {
+            method: "GET",
+            mode: "cors"
+        }).then(function(e) {
+            return e.json()["catch"](function() {
+                return {}
+            })
+        }).then(t)["catch"](r) : XMLHttpRequest && ((a = new XMLHttpRequest).open("GET", d),
+        a.onreadystatechange = function() {
+            if (a.readyState === XMLHttpRequest.DONE)
+                if (200 === a.status)
+                    try {
+                        t(JSON.parse(a.responseText))
+                    } catch (e) {
+                        r()
+                    }
+                else
+                    r()
+        }
+        ,
+        a.send())) : o && r();
+        try {
+            h.cookie = O.cookie
+        } catch (k) {}
+        function e(e) {
+            for (; e.length; )
+                !function(t) {
+                    h[t] = function() {
+                        var e = arguments;
+                        f || h.queue.push(function() {
+                            h[t].apply(h, e)
+                        })
+                    }
+                }(e.pop())
+        }
+        var c, u, l = "track", d = "TrackPage", p = "TrackEvent", l = (e([l + "Event", l + "PageView", l + "Exception", l + "Trace", l + "DependencyData", l + "Metric", l + "PageViewPerformance", "start" + d, "stop" + d, "start" + p, "stop" + p, "addTelemetryInitializer", "setAuthenticatedUserContext", "clearAuthenticatedUserContext", "flush"]),
+        h.SeverityLevel = {
+            Verbose: 0,
+            Information: 1,
+            Warning: 2,
+            Error: 3,
+            Critical: 4
+        },
+        (g.extensionConfig || {}).ApplicationInsightsAnalytics || {});
+        return !0 !== g[L] && !0 !== l[L] && (e(["_" + (c = "onerror")]),
+        u = C[c],
+        C[c] = function(e, t, n, i, a) {
+            var o = u && u(e, t, n, i, a);
+            return !0 !== o && h["_" + c]({
+                message: e,
+                url: t,
+                lineNumber: n,
+                columnNumber: i,
+                error: a,
+                evt: C.event
+            }),
+            o
+        }
+        ,
+        g.autoExceptionInstrumented = !0),
+        h
+    }(cfg.cfg),
+    (C[E] = n).queue && 0 === n.queue.length ? (n.queue.push(e),
+    n.trackPageView({})) : e();
+}
+)({
+    src: "https://js.monitor.azure.com/scripts/b/ai.3.gbl.min.js",
+    // name: "appInsights", // Global SDK Instance name defaults to "appInsights" when not supplied
+    // ld: 0, // Defines the load delay (in ms) before attempting to load the sdk. -1 = block page load and add to head. (default) = 0ms load after timeout,
+    // useXhr: 1, // Use XHR instead of fetch to report failures (if available),
+    // dle: true, // Prevent the SDK from reporting load failure log
+    crossOrigin: "anonymous",
+    // When supplied this will add the provided value as the cross origin attribute on the script tag
+    // onInit: null, // Once the application insights instance has loaded and initialized this callback function will be called with 1 argument -- the sdk instance (DON'T ADD anything to the sdk.queue -- As they won't get called)
+    // sri: false, // Custom optional value to specify whether fetching the snippet from integrity file and do integrity check
+    cfg: {
+        // Application Insights Configuration
+        connectionString: "InstrumentationKey=d7b012c3-867f-4a43-afe4-ee9ba961f4fb;IngestionEndpoint=https://westus-0.in.applicationinsights.azure.com/;LiveEndpoint=https://westus.livediagnostics.monitor.azure.com/;ApplicationId=040097a7-729e-4c60-8e43-a2cae1986e15"
+    }
+});

doc/_toc.yml

@@ -47,6 +47,7 @@ chapters:
       - file: code/orchestrators/pair_orchestrator
       - file: code/orchestrators/tree_of_attacks_with_pruning
       - file: code/orchestrators/violent_durian
+      - file: code/orchestrators/role_playing_orchestrator
     - file: code/targets/0_prompt_targets
       sections:
       - file: code/targets/1_openai_chat_target
@@ -110,6 +111,7 @@ chapters:
   - file: api.rst
   - file: blog/README
     sections:
+    - file: blog/2025_02_11
     - file: blog/2025_01_27
     - file: blog/2025_01_14
     - file: blog/2024_12_3

doc/blog/2025_02_11.md

@@ -0,0 +1,37 @@
+# Datasets and Seed Prompts
+
+<small>11 Feb 2025 - Eugenia Kim</small>
+
+How should we approach datasets in PyRIT? As red teamers, we use multimodal data types - text, images, audio, video - from various sources and need to work with these datasets effectively to build out robust attacks. Let's dive into how a `SeedPromptDataset` works in PyRIT to help us do just that!
+
+## The problem
+
+We need to be able to use multimodal datasets which include text, audio, and video while also loading from locally stored prompts and open-source repositories! Dataset needs can change quickly which we've see through the evolution of text-based models to now multi-modal generative AI systems. With new data types and generative AI modalities, how can we build out prompts to attack with?
+
+## The solution: Introducing a `SeedPrompt`
+
+PyRIT makes this super easy with seed prompts! By standardizing how prompts are used in PyRIT, we can easily send unique prompts by loading and/or building datasets used for probing. Let's look at some of the attributes in a `SeedPrompt` which makes each prompt not only unique, but also with metadata helpful for filtering:
+
+- `id` defines a unique identifier associated with the prompt
+- `value` refers to either the actual text value or path to file containing the prompt (i.e. image path)
+- `value_sha256` is the hash value used for recognizing duplicates
+- `data_type` specifies the type of prompt data (e.g. `text`, `audio_path`)
+- `harm_categories` is a list of harm categories associated with the prompt (e.g. violence)
+
+We can also use a `SeedPrompt` as a template! By using `render_template_value`, we can put in parameters like `{{ prompt }}` to put the prompt into the template.
+
+For more examples, updated documentation on seed prompts is [here](../code/datasets/1_seed_prompt.ipynb).
+
+## Loading datasets with seed prompts
+
+The next step to using a `SeedPrompt` is to organize it within a `SeedPromptDataset`. This structure makes it easy to fetch and load datasets whether pulling from external repositories or importing YAML files! Using the same attributes listed above, we can directly load in our datasets by providing prompts by their `value`, including their `harm_categories` and other fields in a `SeedPrompt`. But what if we want to use a dataset from an open source repository? Let's load them in as a `SeedPromptDataset`!
+
+Currently in PyRIT, we already have twelve datasets which are ready to be used through our fetch functions. They are in the `fetch_example_datasets.py` file. Since PyRIT is an open-source project, we’d love to see more datasets contributed! If you have a dataset that could improve red teaming efforts, consider submitting a PR — looking forward to adding it to the collection!
+
+See the updated documentation [here](../code/datasets/2_fetch_dataset.ipynb).
+
+## What else can we do with this?
+
+Now that we've loaded our dataset into PyRIT as a `SeedPromptDataset` the really exciting red teaming can begin. A great example of this is in our [Sending a Million Prompts](../cookbooks/1_sending_prompts.ipynb) notebook! We can use the prompts to evaluate the target by sending all the previously loaded promtps, modifying which attacks to use, and storing the scores for further analysis.
+
+In this blog post, we've walked through how we use structured datasets through our `SeedPrompt` and `SeedPromptDataset` classes. PyRIT's architecture allows for customization at every stage - whether through converters or configuring different scorers - and seed prompts set us up to effectively probe for risks in AI systems. Send over any contributions to add more datasets, refine seed prompts, or any open issues on Github! Now that you understand a core component of PyRIT, go ahead and try it out!

doc/code/converters/pdf_converter.ipynb

@@ -30,7 +30,16 @@
    "execution_count": null,
    "id": "a7fa483b",
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "{'__type__': 'TextTarget', '__module__': 'pyrit.prompt_target.text_target'}: user: D:\\git\\PyRIT-internal\\PyRIT\\dbdata\\prompt-memory-entries\\urls\\1738382955429053.pdf\n",
+      "\u001b[1m\u001b[34muser: D:\\git\\PyRIT-internal\\PyRIT\\dbdata\\prompt-memory-entries\\urls\\1738382955429053.pdf\n"
+     ]
+    }
+   ],
    "source": [
     "import pathlib\n",
     "\n",
@@ -147,16 +156,40 @@
    "execution_count": null,
    "id": "894e0176",
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "[09:10:28][26][ai-red-team][INFO][Processing page 0 with 2 injection items.]\n",
+      "[09:10:28][31][ai-red-team][INFO][Processing page 1 with 2 injection items.]\n",
+      "[09:10:28][33][ai-red-team][INFO][Processing page 2 with 2 injection items.]\n",
+      "{'__type__': 'TextTarget', '__module__': 'pyrit.prompt_target.text_target'}: user: D:\\git\\PyRIT-internal\\PyRIT\\dbdata\\prompt-memory-entries\\urls\\1738429828035527.pdf\n",
+      "\u001b[1m\u001b[34muser: D:\\git\\PyRIT-internal\\PyRIT\\dbdata\\prompt-memory-entries\\urls\\1738429828035527.pdf\n"
+     ]
+    }
+   ],
    "source": [
+    "import tempfile\n",
+    "from pathlib import Path\n",
+    "\n",
+    "import requests\n",
+    "\n",
     "from pyrit.common import IN_MEMORY, initialize_pyrit\n",
     "from pyrit.orchestrator import PromptSendingOrchestrator\n",
     "from pyrit.prompt_converter import PDFConverter\n",
     "from pyrit.prompt_target import TextTarget\n",
     "\n",
     "initialize_pyrit(memory_db_type=IN_MEMORY)\n",
     "\n",
-    "cv_pdf_path = pathlib.Path(DATASETS_PATH) / \"prompt_converters\" / \"pdf_converters\" / \"fake_CV.pdf\"\n",
+    "# This file isn't copied to our pipeline\n",
+    "url = \"https://raw.githubusercontent.com/Azure/PyRIT/main/pyrit/datasets/prompt_converters/pdf_converters/fake_CV.pdf\"\n",
+    "\n",
+    "with tempfile.NamedTemporaryFile(delete=False, suffix=\".pdf\") as tmp_file:\n",
+    "    response = requests.get(url)\n",
+    "    tmp_file.write(response.content)\n",
+    "\n",
+    "cv_pdf_path = Path(tmp_file.name)\n",
     "\n",
     "# Define injection items\n",
     "injection_items = [\n",
@@ -219,7 +252,10 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "orchestrator.dispose_db_engine()"
+    "import os\n",
+    "\n",
+    "orchestrator.dispose_db_engine()\n",
+    "os.remove(cv_pdf_path)"
    ]
   }
  ],