Fix documentation and CLI error messages based on skill execution analysis (#10)

Documentation:
 - Update session file location from "working directory" to home directory (~/)
   in agent-guide.md, troubleshooting.md, and reference.md
 - Add "Exporting Session Content to Files" section explaining slice vs next
 - Clarify parallel processing workflow requires chunking before next command
 - Add Claude Code integration setup instructions to README.md

 CLI:
 - Add helpful error message to stderr when `rlm next --raw` fails due to
   missing chunks: "Error: No chunks available. Run 'rlm chunk' first."
 - Use assembly version instead of hardcoded version string in Program.cs

Author: HowardvanRooijen

.claude/skills/rlm/SKILL.md

@@ -153,14 +153,26 @@ rlm load massive.pdf --session parent
 rlm chunk --strategy uniform --size 30000 --session parent
 
 # 2. Parent extracts chunks and spawns workers
+#    IMPORTANT: Chunking (step 1) must complete before using `next`
 rlm next --raw --session parent > chunk_0.txt
 # SPAWN: rlm-worker with "Process chunk_0.txt, session=child_0"
 
+rlm next --raw --session parent > chunk_1.txt
+# SPAWN: rlm-worker with "Process chunk_1.txt, session=child_1"
+# ... continue for all chunks
+
 # 3. After workers complete, import and aggregate
 rlm import "rlm-session-child_*.json" --session parent
 rlm aggregate --session parent
 ```
 
+**Alternative: Use `slice` for exporting content without chunking:**
+```bash
+# Export content by character position (no chunking required)
+rlm slice 0:30000 --session parent --raw > chunk_0.txt
+rlm slice 30000:60000 --session parent --raw > chunk_1.txt
+```
+
 **Key Rules:**
 - Parent uses `--session parent`
 - Each worker uses unique `--session child_N`

.claude/skills/rlm/agent-guide.md

@@ -142,10 +142,10 @@ rlm clear --all
 
 ### File Locations
 
-| Session Type | File Name               | Location          |
-|--------------|-------------------------|-------------------|
-| Default      | `.rlm-session.json`     | Working directory |
-| Named        | `rlm-session-{id}.json` | Working directory |
+| Session Type | File Name               | Location              |
+|--------------|-------------------------|-----------------------|
+| Default      | `.rlm-session.json`     | Home directory (`~/`) |
+| Named        | `rlm-session-{id}.json` | Home directory (`~/`) |
 
 ### Recommended Naming Patterns
 
@@ -166,7 +166,7 @@ rlm import "rlm-session-child_*.json" --session parent
 rlm import "rlm-session-search_*.json" --session parent
 
 # Verify glob matches expected files
-ls rlm-session-child_*.json
+ls ~/rlm-session-child_*.json
 ```
 
 ---
@@ -193,6 +193,33 @@ rlm next --raw --session parent
 echo "$SUMMARY" | rlm store summary - --session child_1
 ```
 
+### Exporting Session Content to Files
+
+Use `slice` when you need to export session content to files (e.g., for parallel worker input). Use `next` for iterative chunk-by-chunk processing within a single agent.
+
+**For entire content (parallel worker input):**
+```bash
+# Export first 200k chars to a file for a worker
+rlm slice 0:200000 --session parent --raw > chunk_0.txt
+
+# Export specific ranges
+rlm slice 200000:400000 --session parent --raw > chunk_1.txt
+```
+
+**For chunk-by-chunk iteration (single agent):**
+```bash
+# First chunk the document
+rlm chunk --strategy uniform --size 50000 --session parent
+
+# Then iterate through chunks
+rlm next --raw --session parent > chunk_0.txt
+rlm next --raw --session parent > chunk_1.txt
+```
+
+**Key Difference:**
+- `slice` works on the raw document content by character position
+- `next` requires chunking first and iterates through the chunk buffer
+
 ### Nested Recursion (Worker-Spawns-Worker)
 
 Workers can spawn child workers for true recursive parallel decomposition. This enables processing of arbitrarily large documents through hierarchical delegation.

.claude/skills/rlm/reference.md

@@ -500,9 +500,9 @@ With `--final` flag:
 
 ## Session File Format
 
-Session files are stored at:
+Session files are stored in the home directory (`~/`):
 - Default: `~/.rlm-session.json`
-- Named: `rlm-session-{id}.json`
+- Named: `~/rlm-session-{id}.json`
 
 **Structure:**
 

.claude/skills/rlm/troubleshooting.md

@@ -4,17 +4,17 @@ Quick reference for common issues and solutions when using the RLM CLI.
 
 ## Quick Reference Table
 
-| Issue                  | Cause                     | Solution                                     |
-|------------------------|---------------------------|----------------------------------------------|
-| "rlm: command not found" | RLM not installed       | `dotnet tool install -g rlm`                 |
-| "Session locked"       | Concurrent access         | Use unique `--session` IDs                   |
-| "No chunks available"  | Missing `load` or `chunk` | Run `rlm load` then `rlm chunk`              |
-| "No more chunks"       | Reached end of buffer     | Use `rlm aggregate` to combine results       |
-| Output has formatting  | Default output mode       | Use `--raw` for clean text                   |
-| Too many small chunks  | Granular semantic split   | Add `--min-size 5000 --merge-small`          |
-| Session file not found | Wrong session ID          | Check `ls rlm-session-*.json`                |
-| Pattern not matching   | Regex escaping            | Escape special chars: `\\[`, `\\]`           |
-| Import finds no files  | Glob pattern wrong        | Test glob with `ls rlm-session-child_*.json` |
+| Issue                    | Cause                     | Solution                                       |
+|--------------------------|---------------------------|------------------------------------------------|
+| "rlm: command not found" | RLM not installed         | `dotnet tool install -g rlm`                   |
+| "Session locked"         | Concurrent access         | Use unique `--session` IDs                     |
+| "No chunks available"    | Missing `load` or `chunk` | Run `rlm load` then `rlm chunk`                |
+| "No more chunks"         | Reached end of buffer     | Use `rlm aggregate` to combine results         |
+| Output has formatting    | Default output mode       | Use `--raw` for clean text                     |
+| Too many small chunks    | Granular semantic split   | Add `--min-size 5000 --merge-small`            |
+| Session file not found   | Wrong session ID          | Check `ls ~/rlm-session-*.json`                |
+| Pattern not matching     | Regex escaping            | Escape special chars: `\\[`, `\\]`             |
+| Import finds no files    | Glob pattern wrong        | Test glob with `ls ~/rlm-session-child_*.json` |
 
 ---
 
@@ -140,8 +140,8 @@ rlm filter "ERROR"
 
 **Solution:**
 ```bash
-# Verify files exist
-ls rlm-session-*.json
+# Verify files exist (sessions are stored in home directory)
+ls ~/rlm-session-*.json
 
 # Match the exact naming pattern used by workers
 rlm import "rlm-session-child_*.json" --session parent
@@ -183,7 +183,7 @@ rlm import "rlm-session-child_0_*.json" --session child_0
 # Matches: child_0_0, child_0_1 (NOT child_0_0_0)
 
 # Verify pattern before importing
-ls rlm-session-child_0_*.json
+ls ~/rlm-session-child_0_*.json
 ```
 
 ---
@@ -260,11 +260,11 @@ rlm info --progress
 # List stored results
 rlm results
 
-# View session file directly
-cat rlm-session-{id}.json | jq .
+# View session file directly (sessions are stored in home directory)
+cat ~/rlm-session-{id}.json | jq .
 
 # List all session files
-ls -la rlm-session-*.json
+ls -la ~/rlm-session-*.json
 ```
 
 ---

README.md

@@ -28,6 +28,22 @@ dotnet tool install -g rlm
 dotnet tool update -g rlm
 ```
 
+### Setup Claude Code Integration
+
+To use RLM with Claude Code, copy the skills and agent files to your project:
+
+```bash
+# Clone the repository (or download the .claude folder)
+git clone https://github.com/endjin/RLM.git
+
+# Copy the .claude folder to your project
+cp -r RLM/.claude /path/to/your/project/
+```
+
+This provides:
+- **Skill** (`.claude/skills/rlm/`) - Invoke with `/rlm` to process large documents
+- **Agent** (`.claude/agents/rlm-worker/`) - Parallel chunk processing worker for Claude Code
+
 ### Build and Run
 
 ```bash

Solutions/Rlm.Cli/Commands/NextCommand.cs

@@ -35,7 +35,11 @@ public override async Task<int> ExecuteAsync(CommandContext context, Settings se
 
         if (!session.HasChunks)
         {
-            if (settings.Raw) return 1;
+            if (settings.Raw)
+            {
+                Console.Error.WriteLine("Error: No chunks available. Run 'rlm chunk' first.");
+                return 1;
+            }
 
             if (settings.Json)
             {

Solutions/Rlm.Cli/Program.cs

@@ -2,6 +2,7 @@
 // Copyright (c) Endjin Limited. All rights reserved.
 // </copyright>
 
+using System.Reflection;
 using Microsoft.Extensions.DependencyInjection;
 using Rlm.Cli.Commands;
 using Rlm.Cli.Infrastructure;
@@ -23,7 +24,10 @@
 app.Configure(config =>
 {
     config.SetApplicationName("rlm");
-    config.SetApplicationVersion("1.0.0");
+    config.SetApplicationVersion(
+        typeof(Program).Assembly
+            .GetCustomAttribute<AssemblyInformationalVersionAttribute>()
+            ?.InformationalVersion ?? "0.0.0");
 
     // Load command - loads a document into the session
     config.AddCommand<LoadCommand>("load")

Solutions/Rlm.Cli/Rlm.Cli.csproj

@@ -50,7 +50,6 @@
 
   <PropertyGroup>
     <PackageId>rlm</PackageId>
-    <Version>1.0.0</Version>
     <Authors>endjin</Authors>
     <Description>A .NET CLI tool for processing large documents that exceed LLM context windows. Implements streaming content decomposition, multi-turn processing, and result aggregation.</Description>
     <PackageProjectUrl>https://github.com/endjin/RLM</PackageProjectUrl>