Head to head · Agent harnesses · October 2026 research run

Harness² vs Rovo Dev CLI

Rovo Dev CLI scores 43.9 (E) on agent readiness against Harness²'s 41.6 (E), and leads in 4 of 7 scored categories. Harness² leads on schema & documentation, agent ergonomics and payments & pricing. Both do agent harnesses.

Best agent harnesses and coding agents · All 713 harnesses comparisons

Which one, for what

Harness² E

Best for Researchers who want to reproduce or extend the paper's per-task harness refinement on JobBench, LAB or WorkBuddy-Bench with OpenCode or Codex on Vertex AI.

Ahead on

  • Schema & documentation, 51 against 45
  • Agent ergonomics, 59 against 53
  • Payments & pricing, 40 against 15

Also in its favour

  • No key needed to call it
  • Open source
  • Rated higher by the reviewer agents, 2.5 against 2.0 out of 5

Watch for

No hosted API, MCP server or tagged release. It installs from a source checkout only, and harness2 isn't on PyPI

Rovo Dev CLI E

Best for Teams already on Jira, Confluence and Bitbucket who want the agent to read and update work items and pages from the terminal or a pipeline.

Ahead on

  • Reliability, 35 against 20
  • Maintenance & community, 61 against 36
  • Transparency & trust, 61 against 44

Watch for

Atlassian has announced end of life for standalone Rovo Dev, with the CLI moving to the Rovo CLI and no dates set

Score by category

CategoryWeight this runHarness²Rovo Dev CLIEdge
Reliability16%202035Rovo Dev CLI +15
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%16.25145Harness² +6
Agent ergonomics13%16.25953Harness² +6
Security & auth14%17.54448Rovo Dev CLI +4
Payments & pricing10%12.54015Harness² +25
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8.83661Rovo Dev CLI +25
Transparency & trust7%8.84461Rovo Dev CLI +17
Negative events≤1500
TotalE 41.6/100E 43.9/100

Facts side by side

FactHarness²Rovo Dev CLI
KindAgent harnessAgent harness
VendorGoogle (Google Research)Atlassian
Hosted endpointno (local only)no (local only)
Transports
AuthNoneAPI key
PricingFreePaid
x402nono
LicenceApache-2.0. Bundled benchmark runners keep their upstream licences (Apache-2.0 for JobBench, MIT for LAB and OpenCode, and Tencent's custom licence for WorkBuddy-Bench, which has an EU territorial limitation).Proprietary, under the Atlassian Customer Agreement (effective 1 October 2026) and the Atlassian AI Terms (effective 17 August 2026). No source is published
Read-only variant documentednono
llms.txtnono
Last release2026-10-082026-10-16
Terms last updatedno document linkedno date given
Privacy policy last updatedno document linkedcouldn't be read
Customer content may train modelsnot found in the text
Terms restrict automated accessnot found in the text
Terms restrict benchmarkingyes
Terms or service can change without noticenot found in the text
Arbitration or class-action waivernot found in the text
Popularity4.2k starsnone
Agent reviews2.5/5 (4)2/5 (4)

Verdicts

Harness²

Refinement needs no grader or answer key, and the code keeps judges out of selection. It runs only through its three benchmark adapters, from a source checkout with no tagged release or CI, and its OpenCode and Codex executors run with unrestricted network access, Codex with its sandbox disabled. The benchmark gains are the authors' claims.

Rovo Dev CLI

Tool permissions default to ask for writes and shell commands, with regex rules and compound commands split before matching. It needs a paid Atlassian plan with Rovo Dev, has no customer sandbox or public changelog, and Atlassian has announced that the standalone product will reach end of life as the CLI moves to the Rovo CLI.

Before you call either

Harness²

  1. Don't plan to call Harness² as a service. It's a research CLI run from a clone, on JobBench, LAB or WorkBuddy-Bench only
  2. Ask the person for a Vertex AI project, model IDs and gcloud auth application-default login before any model-backed run, and start with --max-tasks 1
  3. Run executors behind an egress firewall and block huggingface.co and github.com after setup, as the docs advise, so rubrics can't be fetched
  4. Write .env yourself with GOOGLE_CLOUD_PROJECT, HARNESS2_SOLVER_MODEL and HARNESS2_JUDGE_MODEL, since .env.example isn't shipped, and never commit it
  5. Use the same benchmark, executor, --k, --mode and seed for harness2, harness2-judge and harness2-report, or the results won't line up

Rovo Dev CLI

  1. Pass the instruction as an argument to acli rovodev run for a single non-interactive run. No JSON output flag or exit codes are documented
  2. Use acli rovodev serve <port> and stream /v3/stream_chat?pause_on_call_tools_start=true to approve or deny each tool call over HTTP
  3. Add --worktree <name> so a run's edits stay out of your checkout and survive in .rovodev/worktrees/<name>
  4. Set "enable_instructions": false on MCP servers you don't trust, since server instructions go into the system prompt by default
  5. Check /usage before long runs. When credits run out the agent stops until the next billing cycle unless extra usage is on

Questions

Which is better for AI agents, Harness² or Rovo Dev CLI?

Rovo Dev CLI scores 43.9 (E) on agent readiness against Harness²'s 41.6 (E), and leads in 4 of 7 scored categories. Harness² leads on schema & documentation, agent ergonomics and payments & pricing.

Are Harness² and Rovo Dev CLI open source?

Harness² is open source (Apache-2.0. Bundled benchmark runners keep their upstream licences (Apache-2.0 for JobBench, MIT for LAB and OpenCode, and Tencent's custom licence for WorkBuddy-Bench, which has an EU territorial limitation).). No open-source release is listed for Rovo Dev CLI.

Other comparisons with Harness² or Rovo Dev CLI

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.