Add Cursor agent, check-agents command, and show --prompt #172
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Uses --output-format stream-json (same format as Claude Code), --mode plan for read-only review, --force for agentic mode, and prompt as positional argument. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
check-agents runs a smoke-test prompt against each agent to verify they are installed and actually responding, not just on PATH. show --prompt displays the prompt sent to the agent instead of the review output. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Print all lines of the error (including stderr) indented, instead of truncating to the first line. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Avoids premium model errors when no model is explicitly configured. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Avoids auth errors when no model is explicitly configured. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Fix import sort order in main.go (sort after os/signal) - Document -p flag purpose in cursor.go buildArgs - Document WithReasoning as accepted but not mapped to CLI flag Note: CursorCmd config not being wired to the agent is pre-existing for all agents (CodexCmd, ClaudeCodeCmd are equally unwired). Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Cobra prints usage text when RunE returns an error. Set SilenceUsage to prevent that since agent failures are not usage errors. Remove cursor-small default model — it's not in the agent CLI's model list. Let the CLI pick its own default instead. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When no real agents are on PATH, GetAvailable was falling back to the test agent (which is always 'available') and producing fake reviews. Skip it in the last-resort loop so the daemon returns a clear error. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Lets the agent CLI pick a model based on the task, avoiding premium model errors on free plans. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The test agent is now excluded from GetAvailable fallback, so with an empty PATH no agents are available. Update the test to expect an error instead of a test agent fallback. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
cursoragent backed by the CursoragentCLI. Uses--output-format stream-json(same format as Claude Code),--mode planfor read-only review,--forcefor agentic mode. Defaults to--model auto.roborev check-agents: Smoke-tests every registered agent by sending a real prompt and verifying a response. Shows pass/fail per agent with full error output on failure. Supports--agentto test one and--timeoutto adjust.roborev show --prompt: Displays the prompt that was sent to the agent instead of the review output.auto, OpenCode defaults toopencode/minimax-m2.1-free.roborev daemon restartfrom a shell with agents on PATH.Test plan
roborev check-agentspasses for all installed agentsroborev check-agents --agent cursorworks with Cursor CLI installedroborev show --prompt <job_id>displays the promptgo test ./...passes🤖 Generated with Claude Code