Skip to content

SWE-agent takes a GitHub issue and tries to automatically fix it, using your LM of choice. It can also be employed for offensive cybersecurity or competitive coding challenges. [NeurIPS 2024]

Doc: Split out 'hello world' tutorial; lots of changes

Kilian Liereta357cbc44c2b
diff --git a/docs/about.md b/docs/about.md
new file mode 100644
index 00000000..e69de29b
diff --git a/docs/background/index.md b/docs/background/index.md
index 31e86e83..9c874ef7 100644
--- a/docs/background/index.md
+++ b/docs/background/index.md
@@ -1,9 +1,7 @@
-# Learn more about the project.
+# Project overview
This section of the documentation talks about the architecture and research goals of [SWE-agent](#swe-agent).
-Just want to run SWE-agent? Skip ahead to our [installation notes](../installation/index.md).
-
## SWE-agent <a name="swe-agent"></a>
SWE-agent turns LMs (e.g. GPT-4) into software engineering agents that can fix issues in GitHub repositories.
@@ -33,4 +31,39 @@ For in-depth information, read our [paper](https://arxiv.org/abs/2405.15793). If
year={2024},
url={https://arxiv.org/abs/2405.15793}
}
-```
\ No newline at end of file
+```
+
+## SWE-agent <span class="enigma">EnIGMA</span> <a name="enigma"></a>
+
+!!! warning "Compatibility"
+ SWE-agent <span class="enigma">EnIGMA</span> is currently only available for SWE-agent v0.7.0.
+
+SWE-agent <span class="enigma">EnIGMA</span> adds advanced **offensive cybersecurity capabilities**.
+
+On the [NYU CTF benchmark](https://github.com/NYU-LLM-CTF/LLM_CTF_Database), EnIGMA solves **13.5%** of the capture the flag (CTF) challenges, achieving the state-of-the-art performance on the full test set of 200 challenges, **surpassing previous agents by more than 3x** ([leaderboard](https://enigma-agent.com#results)).
+
+We accomplish our results by extending the [🤖 ACIs](../background/aci.md) concept first introduced in SWE-agent, to the cybersecurity domain. We establish the novel [**:gear: Interactive Agent Tools** (IATs)](iat.md) concept, which enables our agent to use interactive tools such as a debugger, in a multitasking way such that the agent still has access to the main shell while using the debugger.
+
+We also use a new **Summarizer** concept integrated into the agent to deal with long context. Read more about our different summarizers [here](../config/summarizers.md).
+
+Specific [demonstrations](../config/demonstrations.md) were built per each CTF category (cryptography, reverse-engineering, forensics, ...), to enhance the model ability to solve new tasks from the same category.
+
+EnIGMA is built and maintained by researchers from Tel-Aviv University, New York University and Princeton University.
+
+For a quick introduction, watch the following video:
+
+<iframe width="560" height="315" src="https://www.youtube.com/embed/IJxqOsNFiCc?si=xtIxyCcriM9FJexK" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>
+
+For all the details, read our [paper](https://arxiv.org/abs/2409.16165). If you found this work helpful, please consider using the following citation:
+
+```bibtex
+@misc{abramovich2024enigmaenhancedinteractivegenerative,
+ title={EnIGMA: Enhanced Interactive Generative Model Agent for CTF Challenges},
+ author={Talor Abramovich and Meet Udeshi and Minghao Shao and Kilian Lieret and Haoran Xi and Kimberly Milner and Sofija Jancheska and John Yang and Carlos E. Jimenez and Farshad Khorrami and Prashanth Krishnamurthy and Brendan Dolan-Gavitt and Muhammad Shafique and Karthik Narasimhan and Ramesh Karri and Ofir Press},
+ year={2024},
+ eprint={2409.16165},
+ archivePrefix={arXiv},
+ primaryClass={cs.AI},
+ url={https://arxiv.org/abs/2409.16165},
+}
+```
diff --git a/docs/dev/contribute.md b/docs/dev/contribute.md
index 67fbcab4..173f5f3f 100644
--- a/docs/dev/contribute.md
+++ b/docs/dev/contribute.md
@@ -64,7 +64,7 @@ pytest -n auto
If you are using VSCode, you might want to add the following two files:
<details>
-<summary>`.vscode/launch.json`</summary>
+<summary><code>.vscode/launch.json</code></summary>
```json
--8<-- "docs/dev/vscode_launch.json"
@@ -72,7 +72,7 @@ If you are using VSCode, you might want to add the following two files:
</details>
<details>
-<summary>`.vscode/settings.json`</summary>
+<summary><code>.vscode/settings.json</code></summary>
```json
--8<-- "docs/dev/vscode_settings.json"
diff --git a/docs/installation/keys.md b/docs/installation/keys.md
index 10dad215..ee434b31 100644
--- a/docs/installation/keys.md
+++ b/docs/installation/keys.md
@@ -22,3 +22,19 @@ See the following links for tutorials on obtaining [Anthropic](https://docs.anth
!!! tip "Models"
Some more information about the available models in our [usage FAQ](../usage/usage_faq.md).
+
+### Models for testing
+
+We also provide models for testing SWE-agent without spending any credits
+
+* `HumanModel` and `HumanThoughtModel` will prompt for input from the user that stands in for the output of the LM. This can be used to create new [demonstrations](../config/demonstrations.md#manual).
+* `ReplayModel` takes a trajectory as input and "replays it"
+* `InstantEmptySubmitTestModel` will create an empty `reproduce.py` and then submit
+
+### Debugging
+
+* If you get `Error code: 404`, please check your configured keys, in particular
+ whether you set `OPENAI_API_BASE_URL` correctly (if you're not using it, the
+ line should be deleted or commented out).
+ Also see [this issue](https://github.com/princeton-nlp/SWE-agent/issues/467)
+ for reference.
\ No newline at end of file
diff --git a/docs/installation/source.md b/docs/installation/source.md
index 283f359e..3d05550c 100644
--- a/docs/installation/source.md
+++ b/docs/installation/source.md
@@ -30,8 +30,6 @@ Installation from source is the preferred way to set up SWE-agent on your machin
SWE-agent is still in active development. Features and enhancement are added often.
To make sure you are on the latest version, periodically run `git pull`
(there is no need to redo the `pip install`).
- You might also want to run `docker pull sweagent/swe-agent:latest` or `./setup.sh` periodically
- (though changes to the container are more rare).
!!! note "Development setup"
Want to modify SWE-agent? Great! There are a few extra steps and tips:
diff --git a/docs/usage/cl_tutorial.md b/docs/usage/cl_tutorial.md
index 07b3ea4e..1a7a23c0 100644
--- a/docs/usage/cl_tutorial.md
+++ b/docs/usage/cl_tutorial.md
@@ -1,25 +1,18 @@
-# Command line usage tutorial
+# Command line basics
-This tutorial walks you through running SWE-agent from the command line.
-Beginners might also be interested in the our web-based GUI (see [here](web_ui.md)).
-This tutorial focuses on using SWE-agent as a tool to solve individual issues.
-Benchmarking SWE-agent is covered [separately](benchmarking.md).
-Finally, we have a different tutorial for using SWE-agent for [coding challenges](coding_challenges.md).
+!!! abstract "Command line basics"
+ This tutorial walks you through running SWE-agent from the command line.
-## Getting started
+ * Please read our [hello world](hello_world.md) tutorial before proceeding.
+ * This tutorial focuses on using SWE-agent as a tool to solve individual issues.
+ Benchmarking SWE-agent is covered [separately](benchmarking.md).
+ Finally, we have a different tutorial for using SWE-agent for [coding challenges](coding_challenges.md).
-After installing SWE-agent, you have the `sweagent` command available. Run `sweagent --help` to see the list of subcommands.
-The most important ones are
+## A few examples
-* `sweagent run`: Run SWE-agent on a single problem statement.
-* `sweagent run-batch`: Run SWE-agent on a list of problem statements. This is what you would use for benchmarking, or when
- working with a larger set of historic issues.
+Before we start with a more structured explanation of the command line options, here are a few examples that you might find immediately useful:
-In this tutorial, we will focus on the `run` subcommand.
-
-Let's start with an absolutely trivial example and solve an issue about a simple syntax error ([`swe-agent/test-repo #1`](https://github.com/SWE-agent/test-repo/issues/1))
-
-```bash
+```bash title="Fix a github issue"
python run.py \
--agent.model.name=gpt4 \
--agent.model.per_instance_cost_limit=2.00 \ # (1)!
@@ -27,60 +20,14 @@ python run.py \
--problem_statement.github_url=https://github.com/SWE-agent/test-repo/issues/1
```
-1. This limits the inference cost per instance to $2. The default is $3.
-
-!!! tip "Annotations"
- Notice the :material-chevron-right-circle: icon in the left margin in the code snippet? Click on it to display more information
- about the line.
-
-The example above uses the `gpt4` model from OpenAI. In order to use it, you need to add your OpenAI API key to the environment:
-
-```bash
-export OPENAI_API_KEY=<your key>
-```
-
-alternatively, you can create a `.env` file in your working directory and put the key in there.
-More information about environment: TODO
-
-
-TODO: Update output
-
-<details>
-<summary>Output</summary>
-
-```json
---8<-- "docs/usage/cl_tutorial_cmd_1_output.log"
+```bash title="Work on a github repo with a custom problem statement" hl_lines="4"
+python run.py \
+ ...
+ --env.repo.github_url=https://github.com/SWE-agent/test-repo \
+ --problem_statement.text="Hey, can you fix all the bugs?"
```
-</details>
-
-As you can see, the command line options are hierarchical. At the top level, there are three important sections:
-
-* `problem_statement`: What problem are you trying to solve?
-* `agent`: How do you want to solve the problem? This includes setting up the LM with `--agent.model`.
-* `env`: What is the environment in which the problem statement should be solved?
- This includes setting the repository/folder with the source files with `--env.repo`, as well as docker images and other dependencies.
- This will also control where the code is executed (in a local container or in the cloud).
-
-
-Watching the output, you can notice several stages:
-
-1. Setting up the **deployment**: SWE-agent lets LMs execute actions in sandboxed environments. It can run these environments
- in docker containers (default), on modal, AWS fargate, or directly on your computer (not recommended).
- When the deployment starts, you will notice a "starting runtime" message that takes a few seconds. The runtime is
- what is executing the commands within your deployment.
-2. Setting up **tools**: The tools that you specified are copied and installed within the environment.
-3. **System and instance prompts**: The initial instructions are shown to the LM.
-4. **Main loop**: The LM starts to suggest and execute actions.
-5. **Submission**: The LM calls `submit` and we extract the patch (i.e., the changes to the source code that solve the problem).
-
-The complete details of the run are saved as a "trajectory" file (more about them [here](trajectories.md)). They can also be turned into new [demonstrations](../config/demonstrations.md) together with other log and output files.
-
-## A few more examples
-
-Before we continue with a more structured explanation of the command line options, here are some more examples that you might find immediately useful:
-
-```bash title="Local repository with problem statement from file and custom docker image" hl_lines="4 5 6"
+```bash title="Fix a bug in a local repository using a custom docker image" hl_lines="4 5 6"
git clone https://github.com/SWE-agent/test-repo.git
python run.py \
--agent.model.name=claude-3.5 \ # (1)!
@@ -92,12 +39,6 @@ python run.py \
1. Make sure to add anthropic keys to the environment for this one!
2. This points to the [dockerhub image](https://hub.docker.com/_/python) of the same name
-```bash title="Github repo with custom problem statement" hl_lines="4"
-python run.py \
- ...
- --env.repo.github_url=https://github.com/SWE-agent/test-repo \
- --problem_statement.text="Hey, can you fix all the bugs?"
-```
For the next example, we will use a cloud-based execution environment instead of using local docker containers.
For this, you first need to set up a modal account, following the instructions at XXX
@@ -181,7 +122,7 @@ So to make sure that we get the default templates in the above examples with `--
--config config/default.yaml
```
-in addition to all the other `--config` options.
+in addition to all the other `--config` options for the two examples above.
## Specifying the repository
diff --git a/docs/usage/coding_challenges.md b/docs/usage/coding_challenges.md
index 85444b69..6f6820f5 100644
--- a/docs/usage/coding_challenges.md
+++ b/docs/usage/coding_challenges.md
@@ -1,14 +1,14 @@
# Using SWE-agent for coding challenges
-!!! tip "Command line tutorial"
- We also provide a more general [command line tutorial](cl_tutorial.md), which covers
- many more advanced features of SWE-agent, but focuses on its use for software engineering
- problems.
+!!! abstract "Overview"
+ It is easy to use SWE-agent to do more than just software engineering.
+ For example, you can tell SWE-agent to work on leetcode or humaneval-style problems.
-It is easy to use SWE-agent to do more than just software engineering.
-For example, you can tell SWE-agent to work on leetcode or humaneval-style problems.
+ Required reading:
+ Please first skim [hello world](hello_world.md) and [command line basics](cl_tutorial.md) to get familiar with the basics of SWE-agent.
-For this, put the problem you want to solve in a markdown file `problem.md`, for example:
+
+Let's start with a new problem statement. For this, put the problem you want to solve in a markdown file `problem.md`, for example:
<details>
<summary>Example leetcode challenge</summary>
diff --git a/docs/usage/hello_world.md b/docs/usage/hello_world.md
new file mode 100644
index 00000000..a93eeae2
--- /dev/null
+++ b/docs/usage/hello_world.md
@@ -0,0 +1,75 @@
+# Hello world
+
+!!! abstract "Fix a github issue with SWE-agent"
+ In this tutorial, we will fix a github issue with SWE-agent using the command line interface.
+
+ * Make sure you have [installed](../installation/index.md) SWE-agent and have a [language model](../installation/keys.md) set up.
+ * We will be executing code in a Docker sandbox, so make sure you have docker installed ([docker troubleshooting](../installation/tips.md)).
+ If you cannot run docker, skim this tutorial and see how you can run SWE-agent with cloud-based execution in the [command line basics tutorial](cl_tutorial.md).
+
+After installing SWE-agent, you have the `sweagent` command available. Run `sweagent --help` to see the list of subcommands.
+The most important ones are
+
+* `sweagent run`: Run SWE-agent on a single problem statement.
+* `sweagent run-batch`: Run SWE-agent on a list of problem statements. This is what you would use for benchmarking, or when
+ working with a larger set of historic issues.
+
+In this tutorial, we will focus on the `run` subcommand.
+
+Let's start with an absolutely trivial example and solve an issue about a simple syntax error ([`swe-agent/test-repo #1`](https://github.com/SWE-agent/test-repo/issues/1))
+
+```bash
+python run.py \
+ --agent.model.name=gpt4 \
+ --agent.model.per_instance_cost_limit=2.00 \ # (1)!
+ --env.repo.github_url=https://github.com/SWE-agent/test-repo \
+ --problem_statement.github_url=https://github.com/SWE-agent/test-repo/issues/1
+```
+
+1. This limits the inference cost per instance to $2. The default is $3.
+
+!!! tip "Annotations"
+ Notice the :material-chevron-right-circle: icon in the left margin in the code snippet? Click on it to display more information
+ about the line.
+
+The example above uses the `gpt4` model from OpenAI. In order to use it, you need to add your OpenAI API key to the environment:
+
+```bash
+export OPENAI_API_KEY=<your key>
+```
+
+alternatively, you can create a `.env` file in your working directory and put the key in there.
+More information about environment: TODO
+
+
+TODO: Update output
+
+<details>
+<summary>Output</summary>
+
+```json
+--8<-- "docs/usage/cl_tutorial_cmd_1_output.log"
+```
+</details>
+
+As you can see, the command line options are hierarchical. At the top level, there are three important sections:
+
+* `problem_statement`: What problem are you trying to solve?
+* `agent`: How do you want to solve the problem? This includes setting up the LM with `--agent.model`.
+* `env`: What is the environment in which the problem statement should be solved?
+ This includes setting the repository/folder with the source files with `--env.repo`, as well as docker images and other dependencies.
+ This will also control where the code is executed (in a local container or in the cloud).
+
+
+Watching the output, you can notice several stages:
+
+1. Setting up the **deployment**: SWE-agent lets LMs execute actions in sandboxed environments. It can run these environments
+ in docker containers (default), on modal, AWS fargate, or directly on your computer (not recommended).
+ When the deployment starts, you will notice a "starting runtime" message that takes a few seconds. The runtime is
+ what is executing the commands within your deployment.
+2. Setting up **tools**: The tools that you specified are copied and installed within the environment.
+3. **System and instance prompts**: The initial instructions are shown to the LM.
+4. **Main loop**: The LM starts to suggest and execute actions.
+5. **Submission**: The LM calls `submit` and we extract the patch (i.e., the changes to the source code that solve the problem).
+
+The complete details of the run are saved as a "trajectory" file (more about them [here](trajectories.md)). They can also be turned into new [demonstrations](../config/demonstrations.md) together with other log and output files.
\ No newline at end of file
diff --git a/docs/usage/inspector.md b/docs/usage/inspector.md
index 08ac5a76..b7a9ed90 100644
--- a/docs/usage/inspector.md
+++ b/docs/usage/inspector.md
@@ -1,10 +1,19 @@
-# Trajectory inspector
+# Inspecting trajectories
-We provide a web interface for visualizing [`.traj` files](trajectories.md) from the `trajectories` folder more easily.
+!!! abstract "Inspecting trajectories"
+ * Trajectories are the main output of SWE-agent. They are the best way to understand what SWE-agent does, especially when running on many problem instances.
+ * We provide a web interface for visualizing [`.traj` files](trajectories.md) from the `trajectories` folder more easily.
+ * Please complete the [hello world](hello_world.md) tutorial before proceeding.
-**Example Usage**
+You should see a folder called `trajectories` in your working directory. Let's go to one of the *experiment directories*:
-Run the inspector in the current directory (this is where your `*.traj` files are):
+```bash
+cd trajectories/$USER/<some directory> # (1)!
+```
+
+1. Don't have a folder here? Make sure to run SWE-agent at least once.
+
+Run the inspector in this directory (this is where your `*.traj` files are):
```bash
sweagent inspector
@@ -14,16 +23,20 @@ The inspector will then be launched in the browser:
![trajectory inspector](../assets/inspector_1.png){: style="width: 49%;"}
![trajectory inspector](../assets/inspector_2.png){: style="width: 49%;"}
-!!! tip
- If you do not see evaluation results, make sure that the SWE-bench output
- is called `results.json` and is in the same directory as the trajectories.
-
-!!! tip
- To see gold patches, point `--data_path` to the SWE-bench dataset.
-
**Additional flags**
- `--directory`: Directory of trajectories to inspect (Defaults to current directory)
- `--port`: Port to host web app (Defaults to `8000`).
+## Benchmark results
+
+If you are running SWE-agent on a benchmark (see [benchmarking](benchmarking.md)), you will see evaluation results as ✅ or ❌.
+Otherwise, you will see ❓.
+
+!!! tip
+ * If you do not see evaluation results, make sure that the SWE-bench output
+ is called `results.json` and is in the same directory as the trajectories.
+ * To see gold patches, point `--data_path` to the SWE-bench dataset.
+
+
{% include-markdown "../_footer.md" %}
diff --git a/docs/usage/trajectories.md b/docs/usage/trajectories.md
index 764d6f8b..dadc8d37 100644
--- a/docs/usage/trajectories.md
+++ b/docs/usage/trajectories.md
@@ -2,8 +2,24 @@
The [`trajectories/`](https://github.com/princeton-nlp/SWE-agent/tree/main/trajectories) folder is the default location that experiment results (for example invocations of [`sweagent run`](cl_tutorial.md)) will be written to.
+## Trajectories
+
The main output file is `<instance_id>.traj`, which is a `.json` formatted file containing the (thought, action, observation) turns generated by SWE-agent towards solving `<instance_id>`.
+Under the `trajectory` key, you can see information for every step of the agent. It looks something like this:
+
+```json
+{
+ "action": "ls -F\n",
+ "observation": "AUTHORS.rst\nCHANGELOG.rst\nCODE_OF_CONDUCT.md...",
+ "response": "We are indeed seeing the same output as the issue. The issue suggests that we should look at line 1474 of the `fields.py`...",
+ "state": "{\"open_file\": \"/marshmallow-code__marshmallow/reproduce.py\", \"working_dir\": \"/marshmallow-code__marshmallow\"}\n",
+ "thought": "We are indeed seeing the same output as the issue. The issue suggests that we should look at line 1474 of the `fields.py`..."
+},
+```
+
+Here's a full example:
+
<details>
<summary>Example trajectory</summary>
@@ -14,6 +30,15 @@ Note: This trajectory is from SWE-agent v0.7.0. It might look different in newer
```
</details>
+!!! tip "Viewing trajectories"
+ We provide a [trajectory viewer](inspector.md) for an easy viewing of trajectories.
+ If you want to view the json, we recommend [jsoneditoronline](https://jsoneditoronline.org/),
+ because of the way it handles long lines.
+
+Trajectories can also be turned into custom demonstrations for SWE-agent ([more information](../config/demonstrations.md)).
+
+## Other output files
+
Together with the trajectory, SWE-agent will also generate the following files per instance:
* `config.yaml`: A yaml file containing the configuration used for the experiment. You can directly repeat the run of this
@@ -59,11 +84,8 @@ Here are a few more files that are generated with `sweagent run-batch`:
* `run_batch_exit_statuses.yaml`: A yaml file containing the exit statuses of all instances in the experiment (for this run of `sweagent run-batch`).
-!!! tip "Viewing trajectories"
- We provide a [trajectory viewer](inspector.md) for an easy viewing of trajectories.
-
!!! tip
- * Evaluation is not completed by `run.py`, it is a separate step (see [benchmarking](benchmarking.md))
+ * Evaluation is not completed by `sweagent run-batch`, it is a separate step (see [benchmarking](benchmarking.md))
* `preds.json` can be referenced directly into SWE-bench to run evaluation (see [benchmarking](benchmarking.md))
* Trajectories can be turned into custom demonstrations for SWE-agent ([more information](../config/demonstrations.md)).
diff --git a/docs/usage/usage_faq.md b/docs/usage/usage_faq.md
index dc7183ee..ea1532c5 100644
--- a/docs/usage/usage_faq.md
+++ b/docs/usage/usage_faq.md
@@ -2,51 +2,11 @@
## What models are supported?
-!!! note "Model support"
- Note that SWE-agent is currently unlikely to perform well with small or local models.
+See [models](../installation/keys.md).
-Models are configured in [`models.py`](https://github.com/princeton-nlp/SWE-agent/blob/main/sweagent/agent/models.py) (we're working on giving a complete list of model settings).
+## What's up with all the output files?
-Here are some few examples:
+You're probably most interested in the `*.traj` files, which contain complete records of SWE-agent's thought process and actions. See [output files](trajectories.md) for more information.
-```
-gpt4
-gpt4o
-gpt4-turbo
-claude-2
-claude-opus
-claude-sonnet
-claude-haiku
-claude-sonnet-3.5
-deepseek-coder
-azure:gpt4
-```
-
-### Ollama support
-
-Models served with an ollama server can be used by specifying `--model` with `ollama:model_name` and `--host_url` to point to the url used to serve ollama (`http://localhost:11434` by default). See more details about using ollama [here](https://github.com/ollama/ollama/tree/main/docs).
-
-```bash
-python run.py --model_name ollama:deepseek-coder:6.7b-instruct \
- --host_url http://localhost:11434 \
- --data_path https://github.com/pvlib/pvlib-python/issues/1603 \
- --config_file config/default_from_url.yaml
-```
-
-### Models for testing
-
-We also provide models for testing SWE-agent without spending any credits
-
-* `HumanModel` and `HumanThoughtModel` will prompt for input from the user that stands in for the output of the LM. This can be used to create new [demonstrations](../config/demonstrations.md#manual).
-* `ReplayModel` takes a trajectory as input and "replays it"
-* `InstantEmptySubmitTestModel` will create an empty `reproduce.py` and then submit
-
-### Debugging
-
-* If you get `Error code: 404`, please check your configured keys, in particular
- whether you set `OPENAI_API_BASE_URL` correctly (if you're not using it, the
- line should be deleted or commented out).
- Also see [this issue](https://github.com/princeton-nlp/SWE-agent/issues/467)
- for reference.
{% include-markdown "../_footer.md" %}
diff --git a/mkdocs.yml b/mkdocs.yml
index 44950f35..b13e17c9 100644
--- a/mkdocs.yml
+++ b/mkdocs.yml
@@ -57,21 +57,27 @@ nav:
- "installation/index.md"
- "In browser": "installation/codespaces.md"
- "From source": "installation/source.md"
- - "API keys": "installation/keys.md"
+ - "Models and keys": "installation/keys.md"
- "Troubleshooting": "installation/tips.md"
- "Tutorials":
+ - "Hello world": "usage/hello_world.md"
- "Command line basics": "usage/cl_tutorial.md"
- "Solving coding challenges": "usage/coding_challenges.md"
+ - "Trajectory inspector": "usage/inspector.md"
+ - "Benchmarking": "usage/benchmarking.md"
- "FAQ": "usage/usage_faq.md"
- - User Guide:
+ - User Guides:
- "usage/index.md"
- - "Trajectories": "usage/trajectories.md"
- - "Trajectory inspector": "usage/inspector.md"
- - "Benchmarking": "usage/benchmarking.md"
- - "Config files": "config/config.md"
- - "Demonstrations": "config/demonstrations.md"
- - Development:
+ - Project overview:
+ - "index": "background/index.md"
- "Architecture": "background/architecture.md"
+ - "Agent tools": "background/aci.md"
+ - "Interactive agent tools": "background/iat.md"
+ - "Output files": "usage/trajectories.md"
+ - "Configuration":
+ - "Config files": "config/config.md"
+ - "Demonstrations": "config/demonstrations.md"
+ - Development:
- "Contribution guide": "dev/contribute.md"
- "Formatting conflicts": "dev/formatting_conflicts.md"
- "Changelog": "dev/changelog.md"