For agent-heavy engineering
Run more agents. Review less guesswork.
Teams that let agents build most of the code find the bottleneck moved: to what the agent is told, and to proving what it did. NEWWORK Projects is built around both ends of that loop.
What hurts today
Agents fill gaps with plausible guesses
A vague ticket gets a confident pull request that solves a slightly different problem. The review takes longer than writing it would have.
Context lives in people’s heads
Which file, which service, which edge case. The agent sees the ticket text and nothing else.
Nobody knows what agents cost
Agent work does not show up anywhere: not in time, not in capacity, not in the budget.
How NEWWORK Projects answers it
Export a month of logged hours as CSV
Acceptance checks
4 of 4 provenGiven a manager, when they export March, then the file has one row per person and day
Pull request #381 via AI agentThe totals match the monthly overview
Test: npm test: 14 passed via AI agentA member cannot export someone else’s hours
Pull request #381Test: npm test: 14 passed via AI agent10,000 rows export in under 2 seconds
Test: 10,000 rows in 0.6 s via AI agent
Pull requests
Show the cost center on project cards
Done, not verified yet:
• Pass the last check with evidence
• Merge or close the open pull request
Export a month of logged hours as CSV
Show the cost center on project cards
Totals off by one day in March
The ticket key in the branch, title or description links the pull request.
- 01Ready for agent, or notA readiness score on every ticket and a robot mark on the board. pm-item-start warns the agent itself when the brief is not ready.
- 02One URL for every agentCursor, Claude Code, Codex, VS Code and any MCP client, with the same 30-plus tools and the same rights as the person who runs them.
- 03Evidence instead of “LGTM”Agents pass checks with the pull request, commit or test run attached. Verified done needs every check, merged pull requests and nothing waiting.
- 04Agent hours, visibleAgents run timers over MCP and their time is booked like everyone’s, marked “via AI agent”.
Agents and MCP are on every plan and never metered. Pro adds pull request webhooks and Brainstorm.
Compare plans →Questions
Do agents need their own accounts?
No. An agent uses a person’s token and acts as that person. Its actions show “via AI agent” in the history.
Can several agents work at once?
Yes. Each person can run several agents with their own tokens. Live timers are one per person, so timers show who is accountable.
Give your next agent a brief.
Free for up to 10 people, with MCP for every agent. No credit card.