feat: add generalized cli_args input to action, move CLI docs to CLI.…

.github/workflows/review.yml

@@ -1,4 +1,4 @@
-name: Test OpenHands AI Action
+name: Review Pull Request
 
 on:
   pull_request:
@@ -7,7 +7,7 @@ on:
       - 'action.yml'
 
 jobs:
-  test-action:
+  review-pr:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout repository
@@ -19,9 +19,41 @@ jobs:
         run: git diff ${{ github.event.pull_request.base.sha }} ${{ github.sha }} > pr_diff.txt
 
       - name: Upload PR Diff Artifact
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name: pr-diff
           path: pr_diff.txt
 
+      - name: Validate GitHub Token
+        env:
+          REVIEW_BOT_GITHUB_TOKEN: ${{ secrets.REVIEW_BOT_GITHUB_TOKEN }}
+        shell: bash
+        run: |
+          if [ -z "$REVIEW_BOT_GITHUB_TOKEN" ]; then
+            echo "::error::REVIEW_BOT_GITHUB_TOKEN secret is not set or is empty. Please configure it in repository/organization secrets."
+            exit 1
+          else
+            echo "REVIEW_BOT_GITHUB_TOKEN is set."
+          fi
+
+      - name: Automatically Review Pull Request
+        id: openhands_valid
+        uses: ./
+        with:
+          prompt: |
+            You are a coding expert that reviews pull requests and provides feedback.
+            Please execute the following steps:
+            1. Verify that the environment variable GITHUB_TOKEN is set, if not stop and report an error.
+            2. Review the current pull request ${{ github.event.pull_request.number }}
+            3. if the PR needs code changes,send an overall PR review comment and each individual line-level comment to Github using Github REST API.
+               or else if the PR is overall good without needing any code changes, 
+               just send an overall comment and a PR approval to Github using Github REST API.
+          llm_api_key: ${{ secrets.LLM_API_KEY }}
+          log_all_events: "true"
+          github_token: ${{ secrets.REVIEW_BOT_GITHUB_TOKEN }}
+          additional_env: |
+            {
+              "SELECTED_REPO": "${{ github.repository }}"
+            }
+
 

.openhands/microagents/knowledge/review_pr.md

@@ -0,0 +1,154 @@
+---
+name: review_pr
+type: task
+version: 1.0.1
+agent: CodeActAgent
+triggers:
+  - review pr
+  - review pull request
+  - pr review
+  - code review
+inputs:
+  - name: PR_NUMBER
+    description: 'The number of the pull request to review (e.g., 158)'
+    required: true
+---
+
+# PR Review Task
+
+This task guides you through reviewing pull requests in ${{ SELECTED_REPO }} according to project standards.
+
+## Finding file changes
+When reviewing file changes, instead of diffing with default branch such as `main` please diff against the base branch of the PR.
+
+## Review Steps and Guidelines for Each Step
+
+A high quality code review addresses 3 aspects:
+
+1. is it the appropriate problem (goal)? (Pass 1)
+2. is it the concise and optimal solution? (Pass 2)
+3. is it the high quality implementation? (Pass 3)
+
+For each pass, if needed, you will submit a PR-level comment and multiple line-level suggested changes via Github API.
+
+When everything is done, you will submit a Github pull request review to approve the PR via Github API.
+
+### **Review Pass 1: Is it the appropriate problem (goal)?**:
+- Review the PR description and linked issues, 
+- Summarize the files changes and see if the PR description matches the changes
+- If the PR description is not aligned with the changes, ask the author to clarify the goal of the PR
+- If the PR description is aligned with the changes, proceed based on the understanding of the goal and the overall repository, 
+comment on whether the goal is the right goal. The right goal of a PR should be
+  - aligned with the overall repository goals, 
+  - not duplicating an existing functionality
+  - should leverage the existing code and architecture unless it's intended for a upgrade/replacement/refactor
+  - should not reinvent the wheel: suggest using popular open source libraries if they exist
+  - should not over engineer: suggest using simple solutions if they exist
+  - should not over generalize such as extracting a piece of code logic into a library / component when they are only being used in one place
+  - should have proper sizing (not too big, not too small).
+
+Send *PR-level comment* to ask for clarification on the goal of the PR. ONLY proceed to next Review Pass if the goal is appropriate, otherwise wait for the author to respond.
+
+### **Review Pass 2: Is it the concise and optimal solution?**:
+   - Understand whether the commit is a small change that fixes a bug or a larger change that implements a new feature/component or refactors code.
+   - If it's a small change, skip architectural feedback, proceed to next step.
+   - If it's a larger change, provide feedback on the architectural choices made.
+      - Summarize the key goal and the solution to that key goal implemented in the PR
+      - Check if the author provided documentations about major design decisions
+      - Summarize the design decisions made by author
+      - Check completenesss: suggest alternative approaches if applicable, and create a table to compare the pros and cons of each approach
+
+Send *PR-level comment* to ask for clarification on the solution of the PR. ONLY proceed to next Review Pass if the solution is concise and optimal, otherwise wait for the author to respond.
+
+### **Review Pass 3: Is it the high quality implementation?**:
+- Verify changes follow `.coderabbit.yaml` rules
+- Suggest the right implementation if the current implementation is not the right one
+- Suggest improvements to the code structure, naming conventions, and documentation
+- Check for proper error handling
+- Ensure consistent formatting
+- Check completeness: suggest other part of the code that needs to be updated to complete the implementation.
+
+- **Documentation**:
+   - Verify README/docstring updates
+   - Check for new configuration needs
+
+- **Testing**:
+   - Confirm adequate test coverage
+   - If unit tests should be added, suggest the right tests to add
+   - Verify tests pass
+
+Send *line-level suggested change* to ask for fixing the implementation of the PR. ONLY proceed to next Review Pass if the implementation is high quality, otherwise wait for the author to respond.
+
+## Using Github API
+
+Please note: use markdown in the comment or reply for better readability if it's more than a plain text.
+
+### This is how you send *PR-level comment* with Github API
+
+```bash
+curl -L \
+  -X POST \
+  -H "Accept: application/vnd.github+json" \
+  -H "Authorization: Bearer <YOUR-TOKEN>" \
+  -H "X-GitHub-Api-Version: 2022-11-28" \
+  https://api.github.com/repos/OWNER/REPO/pulls/PULL_NUMBER/comments \
+-d '{ "body": "[COMMENT]", "commit_id": "'$(git rev-parse HEAD)'", "path": "[FILE_PATH]", "start_line": [START_LINE_NUMBER],"start_side":"[START_SIDE]","line": [LINE_NUMBER],"side":"[SIDE]"}'
+```
+
+Note: even for PR-level comment, you still need to provide the `commit_id`, `path`, and `line` to specify the comment location so that the comment will be shown in the PR page in the most relevant location.
+
+### This is how you send a *line-level suggested change* with github API
+
+```bash
+curl -L \
+  -X POST \
+  -H "Accept: application/vnd.github+json" \
+  -H "Authorization: Bearer <YOUR-TOKEN>" \
+  -H "X-GitHub-Api-Version: 2022-11-28" \
+  https://api.github.com/repos/OWNER/REPO/pulls/PULL_NUMBER/comments \
+-d '{ "body": "suggestion\n[SUGGESTED_CODE_WITH_ESCAPED_NEWLINES]\n", "commit_id": "'$(git rev-parse HEAD)'", "path": "[FILE_PATH]", "start_line": [START_LINE_NUMBER],"start_side":"[START_SIDE]","line": [LINE_NUMBER],"side":"[SIDE]"}'
+```
+
+### This is how you make a reply to a review comment
+
+```bash
+curl -L \
+  -X POST \
+  -H "Accept: application/vnd.github+json" \
+  -H "Authorization: Bearer <YOUR-TOKEN>" \
+  -H "X-GitHub-Api-Version: 2022-11-28" \
+  https://api.github.com/repos/OWNER/REPO/pulls/PULL_NUMBER/comments/COMMENT_ID/replies \
+  -d '{"body":"[REPLY]"}'
+```
+
+### This is how you request changes to a PR
+
+```bash
+# Example, use your own data
+curl -L \
+  -X POST \
+  -H "Accept: application/vnd.github+json" \
+  -H "Authorization: Bearer <YOUR-TOKEN>" \
+  -H "X-GitHub-Api-Version: 2022-11-28" \
+  https://api.github.com/repos/OWNER/REPO/pulls/PULL_NUMBER/reviews \
+  -d '{"commit_id":"ecdd80bb57125d7ba9641ffaa4d7d2c19d3f3091","body":"This is close to perfect! Please address the suggested inline change.","event":"REQUEST_CHANGES","comments":[{"path":"file.md","position":6,"body":"Please add more information here, and fix this typo."}]}'
+```
+
+### This is how you approve a PR
+
+```bash
+# Example, use your own data
+curl -L \
+  -X POST \
+  -H "Accept: application/vnd.github+json" \
+  -H "Authorization: Bearer <YOUR-TOKEN>" \
+  -H "X-GitHub-Api-Version: 2022-11-28" \
+  https://api.github.com/repos/OWNER/REPO/pulls/PULL_NUMBER/reviews \
+  -d '{"commit_id":"ecdd80bb57125d7ba9641ffaa4d7d2c19d3f3091","body":"Great stuff!","event":"APPROVE","comments":[...]}'
+```
+
+## Reference Documents
+- `./coderabbit.yaml` for review standards
+- [Github API: Create a review comment](https://docs.github.com/en/rest/pulls/comments?apiVersion=2022-11-28#create-a-review-comment)
+- [Github API: Create a reply for a review comment](https://docs.github.com/en/rest/pulls/comments?apiVersion=2022-11-28#create-a-reply-for-a-review-comment)
+- [Create a review](https://docs.github.com/en/rest/pulls/reviews?apiVersion=2022-11-28#create-a-review)

.openhands/microagents/repo.md

@@ -0,0 +1,11 @@
+---
+name: repo
+type: repo
+agent: CodeActAgent
+---
+
+This repo contains the code for the Namefi Astra project. It is a web application that allows users to manage their domains and subdomains. Read ./README.md for more information.
+
+## Interacting with Github: use the Github API and GITHUB_TOKEN
+
+When interacting with github, use the github API and GITHUB_TOKEN. Do NOT use the github CLI and do NOT navigate to github.com with the web browser.

.openhands/setup.sh

@@ -0,0 +1,3 @@
+#!/bin/bash
+git config --global user.name "D3Serve Dev Team (via OpenHands)"
+git config --global user.email "[email protected]"

CLI.md

@@ -0,0 +1,63 @@
+# OpenHands Docker CLI Usage
+
+You can run the OpenHands agent via Docker with the following command and options:
+
+```sh
+# Example Docker run command
+
+docker run -it \
+    --pull=always \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.33-nikolaik \
+    -e WORKSPACE_MOUNT_PATH=$WORKSPACE_BASE \
+    -e LLM_API_KEY=$LLM_API_KEY \
+    -e LLM_MODEL=$LLM_MODEL \
+    -e LOG_ALL_EVENTS=true \
+    -v $WORKSPACE_BASE:/opt/workspace_base \
+    -v /var/run/docker.sock:/var/run/docker.sock \
+    -v ~/.openhands-state:/.openhands-state \
+    --add-host host.docker.internal:host-gateway \
+    --name openhands-app-$(date +%Y%m%d%H%M%S) \
+    docker.all-hands.dev/all-hands-ai/openhands:0.33 \
+    python -m openhands.core.main --help
+```
+
+## OpenHands CLI Flags
+
+```
+usage: main.py [-h] [-v] [--config-file CONFIG_FILE] [-d DIRECTORY] [-t TASK] [-f FILE] [-c AGENT_CLS] [-i MAX_ITERATIONS] [-b MAX_BUDGET_PER_TASK] [--eval-output-dir EVAL_OUTPUT_DIR] [--eval-n-limit EVAL_N_LIMIT] [--eval-num-workers EVAL_NUM_WORKERS]
+               [--eval-note EVAL_NOTE] [-l LLM_CONFIG] [--agent-config AGENT_CONFIG] [-n NAME] [--eval-ids EVAL_IDS] [--no-auto-continue] [--selected-repo SELECTED_REPO]
+
+Run the agent via CLI
+
+options:
+  -h, --help            show this help message and exit
+  -v, --version         Show version information
+  --config-file CONFIG_FILE
+                        Path to the config file (default: config.toml in the current directory)
+  -d DIRECTORY, --directory DIRECTORY
+                        The working directory for the agent
+  -t TASK, --task TASK  The task for the agent to perform
+  -f FILE, --file FILE  Path to a file containing the task. Overrides -t if both are provided.
+  -c AGENT_CLS, --agent-cls AGENT_CLS
+                        Name of the default agent to use
+  -i MAX_ITERATIONS, --max-iterations MAX_ITERATIONS
+                        The maximum number of iterations to run the agent
+  -b MAX_BUDGET_PER_TASK, --max-budget-per-task MAX_BUDGET_PER_TASK
+                        The maximum budget allowed per task, beyond which the agent will stop.
+  --eval-output-dir EVAL_OUTPUT_DIR
+                        The directory to save evaluation output
+  --eval-n-limit EVAL_N_LIMIT
+                        The number of instances to evaluate
+  --eval-num-workers EVAL_NUM_WORKERS
+                        The number of workers to use for evaluation
+  --eval-note EVAL_NOTE
+                        The note to add to the evaluation directory
+  -l LLM_CONFIG, --llm-config LLM_CONFIG
+                        Replace default LLM ([llm] section in config.toml) config with the specified LLM config, e.g. "llama3" for [llm.llama3] section in config.toml
+  --agent-config AGENT_CONFIG
+                        Replace default Agent ([agent] section in config.toml) config with the specified Agent config, e.g. "CodeAct" for [agent.CodeAct] section in config.toml
+  -n NAME, --name NAME  Session name
+  --eval-ids EVAL_IDS   The comma-separated list (in quotes) of IDs of the instances to evaluate
+  --no-auto-continue    Disable auto-continue responses in headless mode (i.e. headless will read from stdin instead of auto-continuing)
+  --selected-repo SELECTED_REPO
+                        GitHub repository to clone (format: owner/repo) 

action.yml

@@ -39,6 +39,14 @@ inputs:
       These will be available in the container as environment variables.
     required: false
     default: '{}'
+  cli_args:
+    description: 'Additional CLI arguments to pass to the OpenHands agent'
+    required: false
+    default: ''
+  github_token:
+    description: 'GitHub token for API access (will be passed as GITHUB_TOKEN env var)'
+    required: false
+    default: ''
 
 runs:
   using: 'composite'
@@ -64,6 +72,8 @@ runs:
         LOG_ALL_EVENTS: ${{ inputs.log_all_events }}
         ADDITIONAL_ENV: ${{ inputs.additional_env }}
         PROMPT: ${{ inputs.prompt }}
+        CLI_ARGS: ${{ inputs.cli_args }}
+        GITHUB_TOKEN: ${{ inputs.github_token }}
       run: |
         # Parse additional environment variables from JSON
         if [ "$ADDITIONAL_ENV" != "{}" ]; then
@@ -75,12 +85,21 @@ runs:
           echo "No additional environment variables provided"
         fi
 
+        # Add GITHUB_TOKEN to EXTRA_ENV_ARGS if provided
+        if [ ! -z "$GITHUB_TOKEN" ]; then
+          EXTRA_ENV_ARGS="$EXTRA_ENV_ARGS -e GITHUB_TOKEN=$GITHUB_TOKEN"
+          echo "GITHUB_TOKEN will be passed to the container."
+        else
+          echo "GITHUB_TOKEN was not provided or is empty, not adding to EXTRA_ENV_ARGS."
+        fi
+
         # Run Docker command with all environment variables
+        echo "Final EXTRA_ENV_ARGS (including GITHUB_TOKEN if set): $EXTRA_ENV_ARGS"
         echo "Running OpenHands with the following environment variables:"
         echo "- LLM_MODEL: $LLM_MODEL"
         echo "- LOG_ALL_EVENTS: $LOG_ALL_EVENTS"
-        echo "- Additional variables: $EXTRA_ENV_ARGS"
         echo "- Prompt: $PROMPT"
+        echo "- CLI Arguments: $CLI_ARGS"
 
         # Execute the task
         docker run -i \
@@ -95,4 +114,4 @@ runs:
           -v /var/run/docker.sock:/var/run/docker.sock \
           --add-host host.docker.internal:host-gateway \
           ${{ inputs.openhands_image }} \
-          python -m openhands.core.main -t "${PROMPT}" 
+          python -m openhands.core.main -t "Run export GITHUB_TOKEN=$GITHUB_TOKEN and then ${PROMPT}" $CLI_ARGS