I’m a fan of the Code Mode approach - i.e. letting agents run code rather than narrow functions. Many people agree: CircleCI, Perplexity, etc. In fact, mcp2cli gives MCPs a CLI interface.

I feel the main reason is UNIX composability. I can run CLI commands in a loop, pipe them, etc.

I tested it out on ChatGPT. ChatGPT has an @Gmail Plugin. I build a Local MCP Server that exposes my CLIs, including gws (Google Workspace CLI). I gave it 3 tasks in a single prompt:

  1. Mailbox-scale commitments: Using only @SOURCE, scan my work email from the past 6 months. Find every commitment I made that appears to require follow-up.
  2. Communication analytics: Using only @SOURCE, analyze my work email from the past 12 months. Identify the 20 people I interact with most.
  3. Cross-thread project reconstruction: Using only @SOURCE, reconstruct everything material about $CLIENT from my email over the past year, including relevant attachments.

The @Gmail plugin did surprisingly well. After 19 minutes:

  1. 🟢 It scanned 373 sent messages “containing commitment language”, then checked replies, and gave me a prioritized list of 8 items. Good ones.
  2. 🔴 It said “I can’t do this reliably”, but shared a few clusters of relationthips.
  3. 🟢 It clearly reconstructed the client timeline.

The @LocalMCP plugin had issues. I’m wrapping gws inside a developer MCP plugin, and:

  • ChatGPT has to send large files back and forth from gws via an MCP interface, rather than just read it.
  • ChatGPT’s restrictions didn’t allow it to read the retrieved files - maybe because it was a developer plugin.

Clearly, I know less and mess up more than I think - better to use well-tested well-maintained tools and interfaces.

But anyway, I said: “ChatGPT, use codex on @LocalMCP to process the files from gws.” That took 39 minutes:

  1. 🟢 It scanned far more messages and gave me 10 unresolved items.
  2. 🟢 It managed to scan all contacts and give me the top 20.
  3. 🟢 It clearly reconstructed the client timeline.

The quality and scale of the latter are certainly better.

But my main learnings is: don’t underestimate agents’ ability to loop! They can iterate for long - mode like code than humans.