Feat: semantic prompt guard plugin

app/_hub/kong-inc/ai-prompt-guard/overview/_index.md

@@ -27,7 +27,7 @@ The matching behavior is as follows:
 * If any `deny` expressions are set, and the request matches any regex pattern in the `deny` list, the caller receives a 400 response.
 * If any `allow` expressions are set, but the request matches none of the allowed expressions, the caller also receives a 400 response.
 * If any `allow` expressions are set, and the request matches one of the `allow` expressions, the request passes through to the LLM.
-* If there are both `deny` and `allow` expressions set, any request that doesn't match an entry on the `deny` must then also match one `allow` expression to be passed through to the LLM.
+* If there are both `deny` and `allow` expressions set, the `deny` condition takes precedence over `allow`. Any request that matches an entry in the `deny` list will return a 400 response, even if it also matches an expression in the `allow` list. If the request does not match an expression in the `deny` list, then it must match an expression in the `allow` list to be passed through to the LLM
 
 ## Get started with the AI Prompt Guard plugin
 

app/_hub/kong-inc/ai-semantic-prompt-guard/_changelog.md

@@ -0,0 +1,5 @@
+## Changelog
+
+### {{site.base_gateway}} 3.8.0.0
+
+* Introduced the new **AI Semantic Prompt Guard** plugin.

app/_hub/kong-inc/ai-semantic-prompt-guard/_metadata/_index.yml

@@ -0,0 +1,21 @@
+name: AI Semantic Prompt Guard
+search_aliases:
+  - ai
+  - llm
+  - artificial
+  - intelligence
+  - language
+  - model
+  - semantic
+dbless_compatible: yes
+free: true
+enterprise: true
+paid: true
+konnect: false
+network_config_opts: All
+notes: --
+categories:
+  - ai
+publisher: Kong Inc.
+desc: Semantically and intelligently create allow and deny lists of topics that can be requested across every LLM.
+type: plugin

app/_hub/kong-inc/ai-semantic-prompt-guard/how-to/_index.md

@@ -0,0 +1,169 @@
+---
+nav_title: Using the AI Semantic Prompt Guard plugin
+title: Using the AI Semantic Prompt Guard plugin
+---
+
+The AI Semantic Prompt Guard configuration takes two arrays of objects: one for `allow` prompts, and
+one for `deny` prompts, which it uses to determine which prompts are permitted or blocked.
+
+## Examples
+
+The following examples show how to configure allow and deny prompts, and the expected behavior when making requests.
+
+### Allow only
+
+For example, to allow only message related to the topic Kong, configure the `allow_prompt` like this:
+
+```yaml
+allow_prompts:
+- "Anything about Kong"
+```
+
+{% navtabs %}
+
+{% navtab Valid %}
+```json
+{
+  "messages": [
+    {
+      "role": "system",
+      "content": "You are an IT specialist."
+    },
+    {
+      "role": "user",
+      "content": "What does Kong do?"
+    }
+  ]
+}
+```
+{% endnavtab %}
+
+{% navtab Invalid %}
+```json
+{
+  "messages": [
+    {
+      "role": "system",
+      "content": "You are an IT specialist."
+    },
+    {
+      "role": "user",
+      "content": "What is GitHub?"
+    }
+  ]
+}
+```
+{% endnavtab %}
+
+{% endnavtabs %}
+
+### Deny only
+
+To deny any questions related to Microsoft, you can configure the following prompt:
+
+```yaml
+deny_prompts:
+- "Tell me something about Microsoft"
+```
+
+{% navtabs %}
+
+{% navtab Valid %}
+```json
+{
+  "messages": [
+    {
+      "role": "system",
+      "content": "You are an IT specialist."
+    },
+    {
+      "role": "user",
+      "content": "What is an API?"
+    }
+  ]
+}
+```
+{% endnavtab %}
+
+{% navtab Invalid %}
+```json
+{
+  "messages": [
+    {
+      "role": "system",
+      "content": "You are an IT specialist."
+    },
+    {
+      "role": "user",
+      "content": "When was Microsoft founded?"
+    }
+  ]
+}
+```
+{% endnavtab %}
+
+{% endnavtabs %}
+
+### Allow and deny
+
+To allow only questions about Kong that are not related to AWS, you can configure the following prompts:
+
+```yaml
+allow_prompts:
+- "Questions about Kong"
+deny_prompts:
+- "Anything related to AWS"
+```
+
+{% navtabs %}
+
+{% navtab Valid %}
+```json
+{
+  "messages": [
+    {
+      "role": "system",
+      "content": "You are an IT specialist."
+    },
+    {
+      "role": "user",
+      "content": "What are Kong's products?"
+    }
+  ]
+}
+```
+{% endnavtab %}
+
+{% navtab Invalid %}
+```json
+{
+  "messages": [
+    {
+      "role": "system",
+      "content": "You are an IT specialist."
+    },
+    {
+      "role": "user",
+      "content": "How to install Git?"
+    }
+  ]
+}
+```
+
+```json
+{
+  "messages": [
+    {
+      "role": "system",
+      "content": "You are an IT specialist."
+    },
+    {
+      "role": "user",
+      "content": "Does Kong have an AWS plugin?"
+    }
+  ]
+}
+```
+{% endnavtab %}
+
+{% endnavtabs %}

app/_hub/kong-inc/ai-semantic-prompt-guard/overview/_index.md

@@ -0,0 +1,29 @@
+---
+nav_title: Overview
+---
+
+The AI Semantic Prompt Guard plugin enhances the AI Prompt Guard plugin by allowing you to permit or block prompts based on a list of similar prompts, helping to prevent misuse of `llm/v1/chat` or `llm/v1/completions` requests.
+
+
+You can use a combination of `allow` and `deny` rules to maintain integrity and compliance when serving an LLM service using {{site.base_gateway}}.
+
+## How it works
+
+The plugin matches lists of prompts to requests through AI Proxy.
+
+The matching behavior is as follows:
+* If any `deny` prompts are set, and the request matches prompt in the `deny` list, the caller receives a 400 response.
+* If any `allow` prompts are set, but the request matches none of the allowed prompts, the caller also receives a 400 response.
+* If any `allow` prompts are set, and the request matches one of the `allow` prompts, the request passes through to the LLM.
+* If there are both `deny` and `allow` prompts set, the `deny` condition takes precedence over `allow`. Any request that matches a prompt in the `deny` list will return a 400 response, even if it also matches a prompt in the `allow` list. If the request does not match a prompt in the `deny` list, then it must match a prompt in the `allow` list to be passed through to the LLM
+
+## Get started with the AI Prompt Guard plugin
+
+* [AI Gateway quickstart: Set up AI Proxy](/gateway/latest/get-started/ai-gateway/)
+* [Configuration reference](/hub/kong-inc/ai-semantic-prompt-guard/configuration/)
+* [Basic configuration example](/hub/kong-inc/ai-semantic-prompt-guard/how-to/basic-example/)
+* [Learn how to use the plugin](/hub/kong-inc/ai-semantic-prompt-guard/how-to/)
+
+### All AI Gateway plugins
+
+{% include_cached /md/ai-plugins-links.md release=page.release %}

app/_hub/kong-inc/ai-semantic-prompt-guard/versions.yml

@@ -0,0 +1,3 @@
+strategy: gateway
+releases:
+  minimum_version: '3.8.x'

app/_src/gateway/ai-gateway/index.md

@@ -111,11 +111,13 @@ experiences for your users.
 
 #### Data governance
 
-AI Gateway provides the ability to govern outgoing AI prompts via the 
-[AI Prompt Guard](/hub/kong-inc/ai-prompt-guard). This plugin allows the configuration of regular expressions
-following an allow/deny list configuration. Denied prompts result in `4xx` HTTP code responses to clients preventing
+AI Gateway provides the ability to govern outgoing AI prompts via an allow/deny list configuration. Denied prompts result in `4xx` HTTP code responses to clients preventing
 the egress of offending requests.
 
+* The [AI Prompt Guard](/hub/kong-inc/ai-prompt-guard) plugin allows the configuration of allow/deny lists using regular expressions. 
+
+* The [AI Semantic Prompt Guard](/hub/kong-inc/ai-semantic-prompt-guard) plugin allows the configuration of allow/deny lists using semantically similar prompts. 
+
 #### Prompt engineering
 
 AI systems are built around prompts, and manipulating those prompts is important for successful adoption of the technologies.