Document CodeRabbit CLI review defaults#4
Conversation
📝 WalkthroughWalkthroughThe CodeRabbit review skill documentation is updated to clarify the default agent review command syntax and expand documented CLI options. The "Review Commands" section now provides a "Default agent review" example using ChangesCLI Documentation Update
Estimated code review effort🎯 1 (Trivial) | ⏱️ ~3 minutes Poem
🚥 Pre-merge checks | ✅ 4✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
✨ Simplify code
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@plugins/coderabbit/skills/coderabbit-review/SKILL.md`:
- Around line 45-56: The example "coderabbit review --agent -t all" redundantly
specifies -t all while the docs already state that `--type` defaults to `all`;
either remove `-t all` from the example so it reads "coderabbit review --agent"
or (if you want explicit clarity) keep the example but add a short parenthetical
like "(-t all shown for clarity; all is the default)" and ensure the explanatory
line that says "`--type` defaults to `all`" remains accurate; update the
surrounding text referencing `--type`/`-t` so there is no contradiction.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Central YAML (base), Organization UI (inherited)
Review profile: CHILL
Plan: Enterprise
Run ID: a2cc0f47-b9e5-4007-9669-47f3e01dd512
📒 Files selected for processing (1)
plugins/coderabbit/skills/coderabbit-review/SKILL.md
📜 Review details
🔇 Additional comments (1)
plugins/coderabbit/skills/coderabbit-review/SKILL.md (1)
57-93: LGTM!The "Common scopes and comparisons" examples are clear and well-structured, demonstrating useful command variations. The Output Handling, Result Format, and Guardrails sections provide comprehensive operational guidance with appropriate detail.
| Default agent review: | ||
|
|
||
| ```bash | ||
| coderabbit review --agent | ||
| coderabbit review --agent -t all | ||
| ``` | ||
|
|
||
| Common narrower scopes: | ||
| CLI defaults to know: | ||
|
|
||
| - `coderabbit review` uses plain text output by default. | ||
| - `--agent` changes output to structured findings for agent workflows. | ||
| - `--type` defaults to `all`, which includes committed, staged, unstaged, and untracked changes. | ||
|
|
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Description: Check CodeRabbit CLI help output to verify --type default value
# Check if coderabbit CLI is available and get help for review command
if command -v coderabbit &> /dev/null; then
echo "=== Checking coderabbit review --help ==="
coderabbit review --help 2>&1 | rg -A 5 -i "type|--type|-t"
else
echo "CodeRabbit CLI not found in this environment"
fiRepository: coderabbitai/codex-plugin
Length of output: 114
🏁 Script executed:
#!/bin/bash
# Read the SKILL.md file to see the exact content at lines 45-56
echo "=== Reading SKILL.md (lines 45-65) ==="
sed -n '45,65p' plugins/coderabbit/skills/coderabbit-review/SKILL.mdRepository: coderabbitai/codex-plugin
Length of output: 711
🏁 Script executed:
#!/bin/bash
# Search for CLI code related to the review command and --type defaults
echo "=== Searching for review command implementation ==="
find . -type f \( -name "*.ts" -o -name "*.js" -o -name "*.md" \) | head -20
echo ""
echo "=== Searching for --type flag definition ==="
rg -i "type.*default|--type" --type ts --type js -A 3 -B 1 | head -50Repository: coderabbitai/codex-plugin
Length of output: 9784
🏁 Script executed:
#!/bin/bash
# Look for any configuration files or additional documentation about CLI defaults
echo "=== Searching for CLI or review-related documentation ==="
fd -e md | grep -i "review\|cli\|command" | head -20Repository: coderabbitai/codex-plugin
Length of output: 1233
Clarify redundancy between default command and documented defaults.
The "Default agent review" command on line 48 explicitly includes -t all, but line 55 documents that --type defaults to all. This creates potential confusion for users about whether the flag needs to be specified.
Consider one of these approaches:
- Remove
-t allfrom line 48 since it's redundant with the default - Add clarifying text such as: "The
-t allflag is shown explicitly for clarity, thoughallis already the default"
📝 Proposed fix to remove redundant flag
-coderabbit review --agent -t all
+coderabbit review --agentThen also update line 60:
-coderabbit review --agent -t all
+coderabbit review --agent
coderabbit review --agent -t committed🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@plugins/coderabbit/skills/coderabbit-review/SKILL.md` around lines 45 - 56,
The example "coderabbit review --agent -t all" redundantly specifies -t all
while the docs already state that `--type` defaults to `all`; either remove `-t
all` from the example so it reads "coderabbit review --agent" or (if you want
explicit clarity) keep the example but add a short parenthetical like "(-t all
shown for clarity; all is the default)" and ensure the explanatory line that
says "`--type` defaults to `all`" remains accurate; update the surrounding text
referencing `--type`/`-t` so there is no contradiction.
1 participant
This pull request updates the documentation for the
coderabbit reviewcommand to clarify its usage, especially when using the--agentflag. The changes make it easier to understand the default behaviors and options for running reviews with different scopes.Documentation improvements for review commands:
coderabbit reviewproduces plain text output by default, while--agentswitches output to structured findings for agent workflows.--type(-t) flag defaults toall, covering committed, staged, unstaged, and untracked changes.Summary by CodeRabbit