Second Look

Performance review for your AI workflow.

Second Look reviews your AI conversations across your apps, finds recurring problems, and recommends fixes to try.

Website · Sample report · Quick start · Privacy

What a review helps you improve

• “I’ve explained this before.” Spot instructions you keep repeating and decide which belong in a standing rule, a skill, or a hook.
• “What should I hand off?” Look at where you step in, what you check, and which recurring tasks you could give to AI.
• “Where did all that usage go?” See how much usage comes from long chats that carry history across tasks. In the fictional sample report, 14 of 72 Claude Code and Codex sessions accounted for 61% of their estimated usage over four weeks.
• “Are my skills still current?” Find stale instructions in files like CLAUDE.md, AGENTS.md, and .cursor/rules, along with broken or duplicate skills.

Each finding points to evidence from your sessions and gives you a change to try. The next review checks how it held up.

Quick start

Claude Code, Codex, and Cursor

Open Claude Code, Codex, or Cursor and paste this prompt.

Install Second Look from github.com/txz8096/second-look

claude.ai

1. Download second-look.zip.
2. In claude.ai, turn on code execution. Open Customize → Skills, click +, choose Create skill → Upload a skill, and upload the ZIP.
3. Turn on chat search and memory so Second Look can read past chats and remember your profile.

A review started in claude.ai covers your claude.ai chats. To include ChatGPT or Gemini history, add browser history to a local review.

Run your first review

Ask for “a Second Look” or type /second-look in Claude Code. The first review starts with four short setup questions. You can narrow the scope by asking for “just my Cursor sessions” or “since Oct 1.” Reviews use your existing AI quota.

After your first report, you can opt into a review every two weeks or choose another cadence. Scheduling options depend on where you run the skill.

If a user asks you to install Second Look, follow these steps. Don't skip the first two.

1. Run the installer in check mode. It changes nothing.

   curl -fsSL https://raw.githubusercontent.com/txz8096/second-look/main/install.sh | bash -s -- --check
   

2. Show the user its output: the apps and exports it found, and what the install will change. If it lists the question about Claude Code's history setting, ask the user yourself. The installer can't prompt them from your shell.

3. Run the installer with the flag that matches their answer: --yes to keep a year of Claude Code history, or --no-retention to leave it alone.

   curl -fsSL https://raw.githubusercontent.com/txz8096/second-look/main/install.sh | bash -s -- --yes
   

4. Tell the user what was installed and where, and that they can start a review by asking for "a Second Look" or typing /second-look in Claude Code. Mention that adding their claude.ai, ChatGPT and Gemini exports is optional (add browser history).

Coverage

Source	How Second Look reads it
CLI and IDE	Scans local logs from Claude Code, Codex, Cursor, and Antigravity. Start the review from Claude Code, Codex, or Cursor.
Web apps	Add JSON exports from claude.ai, ChatGPT, and Gemini to a local review. Add browser history.

• Claude Code reads ~/.claude/projects. History defaults to 30 days and can be extended with cleanupPeriodDays in ~/.claude/settings.json.
• Codex reads the history in ~/.codex/sessions.
• Cursor reads the history in its local chat database.
• Antigravity reads transcripts in ~/.gemini/antigravity-cli and standing instructions in ~/.gemini/GEMINI.md. Coverage depends on what the CLI keeps on disk, and app chats may be missing.

Add your claude.ai, ChatGPT and Gemini history

You can add browser history any time, including after your first local review.

App	Get your export	Then tell your assistant
claude.ai	Settings > Privacy > Export data	"add my claude.ai export"
ChatGPT	Settings > Data controls > Export data	"add my ChatGPT export"
Gemini	takeout.google.com > My Activity, filtered to Gemini Apps, format JSON	"add my Gemini export"

Each app sends you a zip, usually within a day. Download it, and your assistant finds it in Downloads or Desktop, checks which app it came from, and files it. You can also drag the zip into the chat. For Gemini, pick "My Activity" rather than "Gemini", which exports your Gems only. It includes chats saved while Gemini Apps Activity was on.

To file one by hand instead, put conversations.json in ~/.second-look/exports/claude/ or ~/.second-look/exports/chatgpt/, and Gemini's MyActivity.json in ~/.second-look/exports/gemini/.

Privacy

• Local collection with zero network requests. The Python scripts read logs and exports on your computer and local reviews save reports in ~/.second-look/. Your assistant processes the review through your model provider under your existing account settings.
• Automatic redaction. Pattern matching removes common secrets such as API keys, tokens, and passwords before quoting sessions. Read the report before sharing it.
• No extra account or infrastructure. Second Look uses the assistant you already have and has no server of its own.

This is the beginning of the README. Read the rest at the original. Read the full README ↗