Get instant risk analysis from Claude, Codex, and Gemini on every code change. Auto-fix issues with one command, or run all three models in consensus mode.
Free · 5 credits/month · No credit card required
Everything runs from your terminal.
| 2ndopinion | Smart default: analyze if changes exist, or show interactive menu |
| 2ndopinion menu | Open the interactive command hub |
| 2ndopinion analyze | Analyze current changes (staged + unstaged) |
| 2ndopinion fix | Auto-fix risks with interactive review |
| 2ndopinion undo | Revert the last round of auto-fixes |
| 2ndopinion review --pr <N> | Review a GitHub Pull Request |
| 2ndopinion review --post | Post review as a PR comment |
| 2ndopinion watch | Continuous monitoring on file changes |
| 2ndopinion --llm claude|codex|gemini | Choose which LLM to use |
| 2ndopinion --context-depth minimal|standard|full | Control how much context is gathered |
| 2ndopinion --no-context | Skip context gathering (send diff only) |
| 2ndopinion auth github | Connect your GitHub account |
| 2ndopinion signup | Create a new account |
| 2ndopinion login | Log in to your account |
| 2ndopinion status | Check usage and subscription |
| 2ndopinion upgrade | Upgrade your plan |
| 2ndopinion config | Show current configuration |
| 2ndopinion config --llm <name> | Set default LLM |
| 2ndopinion config --consensus | Toggle consensus mode (3-model analysis) |
| 2ndopinion logout | Log out of your account |
| 2ndopinion --quiet | Suppress splash screen (for scripting/CI) |
| 2ndopinion --debug | Show debug info (API requests, timing) |
Review any GitHub Pull Request from your terminal. Fetches the PR diff, runs it through AI analysis, and optionally posts a formatted review comment with risk scores directly on the PR.
Connect with 2ndopinion auth github — available on the Agent plan.
Don't just find risks — fix them. Run 2ndopinion fix to analyze your changes, generate concrete patches, and review each one interactively before it touches your code.
Available on the Agent plan. Undo anytime with 2ndopinion undo.
Query all three models in parallel and get a unified verdict. See where Claude, Codex, and Gemini agree on risks — and where they disagree. Agreements are high-confidence findings. Disagreements highlight edge cases worth investigating.
Enable with 2ndopinion config --consensus — available on the Power plan.
Just run 2ndopinion — one command surfaces everything. Analyze, watch, fix, review PRs, or manage settings. No subcommands to memorize.
Pick from Claude, Codex, or Gemini per request. Set a default or switch on the fly with --llm.
Run 2ndopinion watch in a second terminal. Get automatic analysis every time your files change.
Risks are grouped into Security, Data Integrity, Error Handling, Performance, and Breaking Changes — with file references and fix suggestions.
Review any GitHub PR from your terminal with 2ndopinion review. Post risk scores as comments, set thresholds for CI, and auto-detect PRs from your branch.
Automatically gathers full file content, type definitions, test files, and project config alongside your diff. The LLM sees what your code actually does, not just what changed.
Run 2ndopinion fix to generate and apply patches for detected risks. Review each fix interactively, with automatic backups and one-command undo.
Works with any language, any AI assistant, any git repo.