docs: stabilize website diagrams

website/docs/user-guide/features/batch-processing.md

@@ -100,7 +100,7 @@ In the current implementation, distributions assign a probability to **each indi
 
 All output goes to `data/<run_name>/`:
 
-```
+```text
 data/my_run/
 ├── trajectories.jsonl    # Combined final output (all batches merged)
 ├── batch_0.jsonl         # Individual batch results

website/docs/user-guide/features/context-files.md

@@ -103,7 +103,7 @@ Context files are loaded by `build_context_files_prompt()` in `agent/prompt_buil
 
 The final prompt section looks roughly like:
 
-```
+```text
 # Project Context
 
 The following project context files have been loaded and should be followed:

website/docs/user-guide/features/honcho.md

@@ -207,16 +207,17 @@ honcho: {}
 
 Honcho context is fetched asynchronously to avoid blocking the response path:
 
-```
-Turn N:
-  user message
-    → consume cached context (from previous turn's background fetch)
-    → inject into system prompt (user representation, AI representation, dialectic)
-    → LLM call
-    → response
-    → fire background fetch for next turn
-         → fetch context    ─┐
-         → fetch dialectic  ─┴→ cache for Turn N+1
+```mermaid
+flowchart TD
+    user["User message"] --> cache["Consume cached Honcho context<br/>from the previous turn"]
+    cache --> prompt["Inject user, AI, and dialectic context<br/>into the system prompt"]
+    prompt --> llm["LLM call"]
+    llm --> response["Assistant response"]
+    response --> fetch["Start background fetch for Turn N+1"]
+    fetch --> ctx["Fetch context"]
+    fetch --> dia["Fetch dialectic"]
+    ctx --> next["Cache for the next turn"]
+    dia --> next
 ```
 
 Turn 1 is a cold start (no cache). All subsequent turns consume cached results with zero HTTP latency on the response path. The system prompt on turn 1 uses only static context to preserve prefix cache hits at the LLM provider.

website/docs/user-guide/features/hooks.md

@@ -12,7 +12,7 @@ The hooks system lets you run custom code at key points in the agent lifecycle
 
 Each hook is a directory under `~/.hermes/hooks/` containing two files:
 
-```
+```text
 ~/.hermes/hooks/
 └── my-hook/
     ├── HOOK.yaml      # Declares which events to listen for

website/docs/user-guide/features/rl-training.md

@@ -174,21 +174,17 @@ The training loop:
 
 ## Architecture Diagram
 
-```
-┌─────────────────┐     ┌──────────────────┐     ┌─────────────────┐
-│   Atropos API   │◄────│   Environment    │────►│  OpenAI/sglang  │
-│  (run-api)      │     │  (BaseEnv impl)  │     │  Inference API  │
-│  Port 8000      │     │                  │     │  Port 8001      │
-└────────┬────────┘     └──────────────────┘     └────────┬────────┘
-         │                                                │
-         │  Batches (tokens + scores + logprobs)          │
-         │                                                │
-         ▼                                                │
-┌─────────────────┐                                       │
-│  Tinker Trainer  │◄──────────────────────────────────────┘
-│  (LoRA training) │  Serves inference via FastAPI
-│  + FastAPI       │  Trains via Tinker ServiceClient
-└─────────────────┘
+```mermaid
+flowchart LR
+    api["Atropos API<br/>run-api<br/>port 8000"]
+    env["Environment<br/>BaseEnv implementation"]
+    infer["OpenAI / sglang<br/>inference API<br/>port 8001"]
+    trainer["Tinker Trainer<br/>LoRA training + FastAPI"]
+
+    env <--> api
+    env --> infer
+    api -->|"batches: tokens, scores, logprobs"| trainer
+    trainer -->|"serves inference"| infer
 ```
 
 ## Creating Custom Environments

website/docs/user-guide/features/skills.md

@@ -140,7 +140,7 @@ When a missing value is encountered, Hermes asks for it securely only when the s
 
 ## Skill Directory Structure
 
-```
+```text
 ~/.hermes/skills/                  # Single source of truth
 ├── mlops/                         # Category directory
 │   ├── axolotl/