1. X
  2. Ben Holmes
Log inSign up
Ben Holmes
Warp
15.2K posts
Ben Holmes profile banner
user avatar

Ben Holmes

Warp
@BHolmesDev
The Whiteboard Guy @warpdotdev
Ordering more markers
bholmes.dev
Joined September 2018
998
Following
46.5K
Followers
RepliesRepliesArticlesArticlesMediaMedia

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
  • Pinned
    user avatar
    Ben Holmes
    Warp
    @BHolmesDev
    Aug 17
    These are the 5 skills I use to get better output from Opus 5 and GPT 5.6: /grill-me for research /taste-review for design /vercel-react-best-practices for ReactJS quality /simplify to remove the fluff /test-app for e2e verification Here's how they work 👇
    Image
    00:00
  • user avatar
    Ben Holmes
    Warp
    @BHolmesDev
    7h
    Having an agent watch a PR for CI failures is so, so nice
    Wilson reports that PR #15576 is approved and mergeable. Full CI is running for Linux, macOS, Windows, wasm, clippy, and release builds; the red “Check CI results” status is stale from the draft run and should update when the test matrix completes. If anything fails, Wilson will handle it.
  • user avatar
    Ben Holmes
    Warp
    @BHolmesDev
    11h
    Agent skills should improve themselves. Basic recipe: - Scorer agents review past conversations - Filter scored runs to those that "failed" - Let self improvement agents update your skills youtu.be/tjRSWpFnrZc 👀
  • user avatar
    Ben Holmes
    Warp
    @BHolmesDev
    12h
    Chrome DevTools but for agents?
    user avatar
    Zach Lloyd
    Warp
    @zachlloydtweets
    13h
    We're building a debugger for agents internally. Breaks down costs by agent + subagent tasks, shows agent processes on a timeline, tracks the exact dollar amount spent. Would you use it?
    The display shows a timeline of agent tasks, detailing execution times, costs, and tool call errors, with color-coded markers for different events.
  • user avatar
    Ben Holmes
    Warp
    @BHolmesDev
    12h
    Been seriously impressed adding self-improvement to our agent conversations so far. We've merged 5 self-improvement PRs in the past few days, including this one that caught token burn in our message passing system for orchestration
    A side-by-side code diff for v1/agents/triage/agent.md, showing line 33 changed from "7. Complete and report. Report to the orchestrator: the work item reference, the complexity, and the ambiguity that a spec must resolve, if you found ambiguity." to the same text plus an added sentence: "Then stop: do not re-check your inbox or resend the report while nothing new has arrived. The orchestrator will message you again if it needs more from you." Removed text is highlighted red on the left, added text green on the right, followed by an unchanged "### Reproduction" heading.
    Image
    03:23
    user avatar
    Warp
    @warpdotdev
    14h
    Introducing self-improvement loops. The concept is simple: what if agents could improve Skills by reviewing past conversations? Here's the three step loop: - Score conversations from criteria you define - Isolate failures - Generate skill improvements
Advertisement
Advertisement