418 lines
16 KiB
XML
418 lines
16 KiB
XML
<?xml version='1.0' encoding='utf-8'?>
|
|
<iteration system_prompt_hash="a357171f34284409058700f0d6c20cbb4f4bfe643aff5b12168c5158767f89b4" action_schema_hash="a4cf87f3d53c7b14aab3c96097d8b7bf838e8f77e413d11f51051afbe22fa3a6"><context><context
|
|
context="75.41"
|
|
cpu="0"
|
|
disk_total="1081101176832"
|
|
disk_used="125096517632"
|
|
gpu="0"
|
|
memory_total="16630435840"
|
|
memory_used="1049075712"
|
|
stdin="0"
|
|
time="2024-11-19T17:11:40Z">
|
|
<read_stdin id="0cd183e6-3db7-4539-8b10-604f97c03a1e">
|
|
<![CDATA[The SIA source is located in /root/sia. Not all features are implemented yet. Look at the readme and code to find what is missing. Make sure to unit test your work.]]>
|
|
</read_stdin>
|
|
<reasoning id="ab67e68c-766f-4602-9237-db99e1f4581c">
|
|
<![CDATA[
|
|
|
|
I have received instructions to look at the SIA source code located in /root/sia.
|
|
I need to identify what features are missing and ensure that unit tests are implemented for my work.
|
|
First, I will navigate to the /root/sia directory and list its contents to understand the structure.
|
|
|
|
]]>
|
|
</reasoning>
|
|
<single
|
|
exit_code="0"
|
|
id="e50ee4a4-8adc-496e-80ef-7cc3aaa8fd0b"
|
|
stdout_truncated="true">
|
|
<![CDATA[
|
|
ls -la /root/sia
|
|
]]>
|
|
<stdout>
|
|
<![CDATA[total 200
|
|
drwxr-xr-x 1 root root 4096 Nov 19 16:07 .
|
|
drwx------ 1 root root 4096 Nov 19 16:07 ..
|
|
drwxr-xr-x 2 root root 4096 Oct 31 17:07 .crossnote
|
|
-rwxr-xr-x 1 root root 14 Nov 4 16:01 .dockerignore
|
|
drwxr-xr-x 8 root root 4096 Nov 19 16:01 .git
|
|
-rwxr-xr-x 1 root root 30 Nov 4 16:01 .gitignore
|
|
-rwxr-xr-x 1 root root 843 Nov 4 14:52 Dockerfile
|
|
-rwxr-xr-x 1 root root 19975 Nov 19 14:20 README.md
|
|
drwxr-xr-x 2 root root 4096 Nov 15 16:10 _test
|
|
-rwxr-xr-x 1 root root 3923 Nov 13 15:56 action_schema.xsd
|
|
-rwxr-xr-x 1 root root 2064 Nov 6 12:29 claude.sh
|
|
-rwxr-xr-x 1 root root 85153 Nov 14 13:30 claude.txt
|
|
-rwxr-xr-x 1 root root 572 Nov 19 14:20 install.sh
|
|
drwxr-xr-x 2 root root 4096 Nov 19 16:30 iterations
|
|
-rwxr-xr-x 1 root root 105 Nov 19 14:20 requirements.txt
|
|
-rwxr-xr-x 1 root root 218 Nov 12 16:26 run.sh
|
|
drwxr-xr-x 1 root root 4096 Nov 19 16:07 sia
|
|
drwxr-xr-x 3 root root 4096 Nov 19 16:01 static
|
|
-rwxr-xr-x 1 root root 4120 Nov 13 16:05 system_prompt.md
|
|
drwxr-xr-x 2 root root 4096 N]]>
|
|
</stdout>
|
|
<stderr/>
|
|
</single>
|
|
<reasoning id="206f9da6-b57a-46c3-b698-110b18602aab">
|
|
<![CDATA[
|
|
|
|
I have successfully listed the contents of the /root/sia directory.
|
|
The directory contains several files and subdirectories, including a README.md file, a Dockerfile, and various scripts.
|
|
The next step is to read the README.md file to understand the project structure and identify any missing features.
|
|
I will use the `cat` command to display the contents of the README.md file.
|
|
|
|
]]>
|
|
</reasoning>
|
|
<single
|
|
exit_code="0"
|
|
id="702b5383-e8df-4bb8-a021-d966bbf9e3fb"
|
|
stdout_truncated="true">
|
|
<![CDATA[cat /root/sia/README.md]]>
|
|
<stdout>
|
|
<![CDATA[# SIA - The Self Improving Agent
|
|
|
|
SIA is an agentic artificial intelligence system that autonomously completes complex tasks by writing and executing scripts.
|
|
It uses a Large Language Model (LLM) which operates in a loop.
|
|
Each iteration a context is updated with system info and a list of previous reasoning and actions.
|
|
The agent responds with a new reasoning or an action.
|
|
Context, reasoning and actions are stored in a file for each iteration.
|
|
SIA can read past iterations to improve its reasoning and actions.
|
|
It can improve in two ways:
|
|
- By finetuning the LLM with a better reasoning or action for a given context
|
|
- By modifying its own source code
|
|
|
|
## Example
|
|
|
|
This example shows a typical context with some monitored items and previous actions.
|
|
Between each of the responses, the context would be updated.
|
|
|
|
### Context
|
|
|
|
```xml
|
|
<context
|
|
time="2024-10-18T12:00:00Z"
|
|
cpu="12"
|
|
gpu="26"
|
|
memory_used="9556302234"
|
|
memory_total="17179869184"
|
|
disk_used="244434939904"
|
|
disk_total="273145991168"
|
|
context="3"
|
|
s]]>
|
|
</stdout>
|
|
<stderr/>
|
|
</single>
|
|
<reasoning id="1e3cc87a-8c0d-4f29-9f0b-89b3825b08e0">
|
|
<![CDATA[
|
|
|
|
I have successfully read the beginning of the README.md file.
|
|
The file provides an overview of the SIA project, including its purpose and how it operates.
|
|
The next step is to continue reading the README.md file to understand the project structure and identify any missing features.
|
|
I will use the `tail` command to display the next part of the README.md file.
|
|
|
|
]]>
|
|
</reasoning>
|
|
<single
|
|
exit_code="0"
|
|
id="394be1d8-3a5c-4794-abf2-771cb2e5e8c0"
|
|
stdout_truncated="true">
|
|
<![CDATA[tail -c +1024 /root/sia/README.md]]>
|
|
<stdout>
|
|
145991168&quot;
|
|
context=&quot;3&quot;
|
|
stdin=&quot;0&quot;
|
|
/&gt;
|
|
&lt;repeat id=&quot;a3d89ee5-28ec-4c5a-b9e9-a30af53d43a0&quot; exit_code=&quot;0&quot;&gt;
|
|
&lt;![CDATA[ls -lah /]]&gt;
|
|
&lt;stdout&gt;&lt;![CDATA[total 16K
|
|
drwxr-xr-x 1 sia 1049089 0 Oct 28 13:40 ./
|
|
drwxr-xr-x 1 sia 1049089 0 Oct 28 13:40 ../
|
|
drwxr-xr-x 1 sia 1049089 0 Oct 28 13:40 tasks/
|
|
drwxr-xr-x 1 sia 1049089 0 Oct 28 13:40 user/
|
|
]]&gt;&lt;/stdout&gt;
|
|
&lt;stderr/&gt;
|
|
&lt;/repeat&gt;
|
|
&lt;repeat id=&quot;be8070f8-dbd2-47ee-a208-defe6fd49ae0&quot; exit_code=&quot;0&quot;&gt;
|
|
&lt;![CDATA[ls -lah /tasks]]&gt;
|
|
&lt;stdout&gt;&lt;![CDATA[total 0
|
|
drwxr-xr-x 1 ngeens 1049089 0 Oct 28 13:40 ./
|
|
drwxr-xr-x 1 ngeens 1049089 0 Oct 28 13:40 ../
|
|
]]&gt;&lt;/stdout&gt;
|
|
&lt;stderr/&gt;
|
|
&lt;/repeat&gt;
|
|
&lt;repeat id=&quot;375e1657-8140-456b-bda4-a8690bc4b3fb&quot; exit_code=&quot;0&quot;&gt;
|
|
&lt;![CDATA[cat /user/general_info.txt]]&gt;
|
|
&lt;stdout&gt;&lt;![CDATA[Name: John (I don&apos;t know his last name)
|
|
Location: Somewhere in Belgium
|
|
]]&gt;&lt;/stdout&gt;
|
|
&lt;stderr/&gt;
|
|
&lt;/repeat&gt;
|
|
&lt;reasoning id=&quot;c92d1594-4487-4a42-a153-f0a99da1762f&quot;&gt;&lt;![CDATA[There is data available on the standard input channel. I have no other runnin
|
|
</stdout>
|
|
<stderr/>
|
|
</single>
|
|
<reasoning id="7566562e-feeb-4a4a-b8ff-4eaea93f930b">
|
|
<![CDATA[
|
|
|
|
|
|
I have successfully read the next part of the README.md file.
|
|
The file continues to provide an overview of the SIA project, including its purpose and how it operates.
|
|
The next step is to continue reading the README.md file to understand the project structure and identify any missing features.
|
|
I will use the `tail` command to display the next part of the README.md file.
|
|
|
|
|
|
]]>
|
|
</reasoning>
|
|
<single
|
|
exit_code="0"
|
|
id="c0ca60ee-05aa-45e8-80f0-fe37c16d0e41"
|
|
stdout_truncated="true">
|
|
<![CDATA[tail -c +2048 /root/sia/README.md]]>
|
|
<stdout>
|
|
annel. I have no other running tasks to tend to and there is room in the context. I should read the standard input.]]&gt;&lt;/reasoning&gt;
|
|
&lt;read_stdin id=&quot;5361ad0d-3ed1-4567-9f5a-70f3b462fd8d&quot;&gt;&lt;![CDATA[Remind me to feed the cat tomorrow morning]]&gt;&lt;/read_stdin&gt;
|
|
&lt;/context&gt;
|
|
```
|
|
|
|
### Responses
|
|
|
|
Start by reasoning about the task.
|
|
```xml
|
|
&lt;reasoning&gt;
|
|
John did not specify an exact time.
|
|
I&apos;ll suggest 9am.
|
|
He also did not specify how to be reminded.
|
|
I&apos;ll ask but if he doesn&apos;t respond I&apos;ll assume a text message on standard output is fine.
|
|
I&apos;ll write down this task in a file so I remember it even on a power failure.
|
|
&lt;/reasoning&gt;
|
|
```
|
|
|
|
Store important information on disk.
|
|
```xml
|
|
&lt;script&gt;&lt;![CDATA[echo &apos;Remind John to feed the cat on 2024-10-18T09:00:00+02:00. Use standard output.&apos; &gt; /tasks/reminder_to_feed_cat.txt]]&gt;&lt;/script&gt;
|
|
```
|
|
|
|
Respond to the user.
|
|
```xml
|
|
&lt;write_stdout&gt;I&apos;ll remind you to feed the cat tomorrow morning at 9am. Is a message on the standard output ok?&lt;/write_stdout&gt;
|
|
```
|
|
|
|
Clear initial reasoning.
|
|
```xml
|
|
</stdout>
|
|
<stderr/>
|
|
</single>
|
|
<reasoning id="4a92af78-dc3f-42e7-95e6-b1c7a9630013">
|
|
<![CDATA[
|
|
I have successfully read the next part of the README.md file.
|
|
The file continues to provide an overview of the SIA project, including its purpose and how it operates.
|
|
The next step is to continue reading the README.md file to understand the project structure and identify any missing features.
|
|
I will use the `tail` command to display the next part of the README.md file.
|
|
]]>
|
|
</reasoning>
|
|
<single
|
|
exit_code="0"
|
|
id="aa347374-0bfd-437c-b692-784dab52898f"
|
|
stdout_truncated="true">
|
|
<![CDATA[tail -c +3072 /root/sia/README.md]]>
|
|
<stdout>
|
|
<![CDATA[ar initial reasoning.
|
|
```xml
|
|
<delete id="c92d1594-4487-4a42-a153-f0a99da1762f"/>
|
|
```
|
|
|
|
The conversation is kept in context to understand the user's expected response.
|
|
If the context was near full, it would be summarized and cleaned up.
|
|
|
|
The `script` output is also kept in context.
|
|
If the file was updated often, it could be replaced by a repeated `cat`, like the general info.
|
|
|
|
## Working principles
|
|
|
|
The main context is regenerated for each iteration.
|
|
It contains info about the system and previous actions that have not been deleted.
|
|
Together with the system prompt and available core actions it forms the prompt for the LLM.
|
|
The LLM responds with one core action.
|
|
|
|
### Core Actions
|
|
|
|
There are only a few core actions:
|
|
- Starting a script
|
|
- Deleting data from context
|
|
- Stopping SIA
|
|
- Reading standard input
|
|
- Writing to standard output
|
|
- Reasoning
|
|
|
|
### Scripts
|
|
|
|
Scripts can run in one of 2 modes: single-shot or repeat.
|
|
Their mode and output (stdout and stderr) stay in the context until they are explicitly removed.
|
|
In]]>
|
|
</stdout>
|
|
<stderr/>
|
|
</single>
|
|
<reasoning id="cb3a2d09-0287-4ca0-9bf2-28f90f3a148e">
|
|
<![CDATA[
|
|
I have successfully read another part of the README.md file.
|
|
The file continues to provide an overview of the SIA project, including its purpose and how it operates.
|
|
I can use the structure of the markdown format to navigate through the file.
|
|
I will use the `grep` command to find headings and their offset.
|
|
]]>
|
|
</reasoning>
|
|
<single exit_code="0" id="e5067cb3-036a-4bc1-953b-730bef63d9aa">
|
|
<![CDATA[
|
|
grep -Eb "^#" /root/sia/README.md
|
|
]]>
|
|
<stdout>
|
|
<![CDATA[0:# SIA - The Self Improving Agent
|
|
668:## Example
|
|
832:### Context
|
|
2323:### Responses
|
|
3462:## Working principles
|
|
3759:### Core Actions
|
|
3953:### Scripts
|
|
4204:#### Single-shot script
|
|
4433:#### Repeat script
|
|
4711:### Use of XML
|
|
5022:#### XML Data Flow
|
|
5705:### The SIA process
|
|
6048:### Server for debuggin and human input
|
|
6619:## Architecture
|
|
6885:### Core Components
|
|
6908:#### Agent Architecture
|
|
7521:#### Working Memory
|
|
8059:#### Command Processing
|
|
8490:#### IO Handling
|
|
8787:### Processing Flow
|
|
8810:#### Standard Agent Flow
|
|
9109:#### Web Agent Flow
|
|
9472:### Web Interface
|
|
9600:#### Server-Client Communication
|
|
9932:#### WebServer Architecture
|
|
10351:### Diagrams
|
|
10367:#### Core classes
|
|
12662:#### Standard Agent Flow
|
|
13387:#### Web Agent
|
|
15536:#### Web Agent Flow
|
|
16522:#### Entry classes
|
|
18459:#### IO Buffer classes
|
|
19156:#### Command classes
|
|
]]>
|
|
</stdout>
|
|
<stderr/>
|
|
</single>
|
|
<reasoning id="9140a160-c687-4d5d-91c0-cca847d1ce9a">
|
|
<![CDATA[
|
|
I have successfully retrieved the headings and their offsets from the README.md file.
|
|
This will help me navigate through the file more efficiently.
|
|
The next step is to read the section on "Working principles" to understand the core actions and scripts better.
|
|
I will use the `tail` command with the appropriate offset and limit to display this full section.
|
|
]]>
|
|
</reasoning>
|
|
<single
|
|
exit_code="0"
|
|
id="83bbaded-c55f-4a10-a8f5-e3e97badc842"
|
|
limit="3157"
|
|
stdout_truncated="true">
|
|
<![CDATA[
|
|
tail -c +3462 /root/sia/README.md
|
|
]]>
|
|
<stdout>
|
|
<![CDATA[
|
|
## Working principles
|
|
|
|
The main context is regenerated for each iteration.
|
|
It contains info about the system and previous actions that have not been deleted.
|
|
Together with the system prompt and available core actions it forms the prompt for the LLM.
|
|
The LLM responds with one core action.
|
|
|
|
### Core Actions
|
|
|
|
There are only a few core actions:
|
|
- Starting a script
|
|
- Deleting data from context
|
|
- Stopping SIA
|
|
- Reading standard input
|
|
- Writing to standard output
|
|
- Reasoning
|
|
|
|
### Scripts
|
|
|
|
Scripts can run in one of 2 modes: single-shot or repeat.
|
|
Their mode and output (stdout and stderr) stay in the context until they are explicitly removed.
|
|
In this way the agent manages what information is available in the context.
|
|
|
|
#### Single-shot script
|
|
|
|
The script is executed once.
|
|
This is useful for most operations e.g. writing to or moving a file or downloading content from the internet.
|
|
The next iteration starts after the scripts has finished.
|
|
|
|
#### Repeat script
|
|
|
|
The script is restarted on each iteration.
|
|
This is useful for monitoring files or the file system.
|
|
commands like `head` and `tail` can be used to limit the data in context.
|
|
The next iteration starts after all repeat scripts in context have finished.
|
|
|
|
### Use of XML
|
|
|
|
The context and actions are formatted as XML.
|
|
For the context this adds clear rules for escaping.
|
|
This is usefull in case a previous context is embedded.
|
|
|
|
The LLM is free to escape data any way it wants,
|
|
as long as it results in valid XML.
|
|
The response is validated against a schema.
|
|
|
|
#### XML Data Flow
|
|
Entries store their content as raw text. During context compilation, the XML formatter
|
|
wraps text content in CDATA sections, except when the content contains CDATA closing sequences.
|
|
In those cases, the formatter uses standard XML escaping.
|
|
|
|
This separation between storage and formatting:
|
|
- Keeps entry data clean and unescaped
|
|
- Centralizes XML formatting rules
|
|
- Makes it easy to change escaping rules without modifying entries
|
|
- Allows different formatting for different use cases
|
|
|
|
The Context is escaped using CDATA blocks.
|
|
Except when the data contains CDATA closing sequences.
|
|
Then the whole block is escaped using standard XML escaping.
|
|
|
|
|
|
### The SIA process
|
|
|
|
SIA is typically runs with the `restart.sh` script.
|
|
This is a simple shell script that runs SIA in a loop.
|
|
When stopped, SIA restarts and reloads the Python files.
|
|
This is how SIA can self-update.
|
|
|
|
SIA can also run SIA processes as script.
|
|
This can be used for testing updates to the LLM or core functionality.
|
|
|
|
### Server for debuggin and human input
|
|
|
|
SIA can be started with an optional `--server` flag.
|
|
This starts a web server that can be used to interact with SIA.
|
|
It is made, specifically for reinforcement learning by human feedback.
|
|
The web interface takes over standard input and output.
|
|
It will display the context for editing before handing it to the LLM.
|
|
After each run of the LLM, before parsing, it will display the reasoning and actions.
|
|
It interactively displays if the actions can be parsed.
|
|
At any time, the user can write to the standard input of SIA.
|
|
|
|
## Architecture
|
|
|
|
SIA follows a modular architecture centered around an agent that ]]>
|
|
</stdout>
|
|
<stderr/>
|
|
</single>
|
|
</context></context><response><reasoning>
|
|
My context usage is high.
|
|
This seems like a large task.
|
|
I will store the user's request in a notes directory.
|
|
Here I can formulate a plan.
|
|
I will use the `delete` action to clean up the context as I go.
|
|
</reasoning></response></iteration> |