// verdict · checked 2026-08-11 · v0 (v0.app docs: agentic features, terminal commands)
v0
Vercel
One prompt in, one gorgeous UI out. Beautiful. Not an agent. Doesn't claim to be — refreshing.
- CAPABILITY
- 5/5 AGENT
- PROOF
- DOCS CLAIMEDThe vendor's documentation states it and we read the page; we did not watch it happen
- TRACE ACCESS
- VENDOR-ONLYSteps appear in a demo or worked example you cannot run
- LEASH
- Not measuredno execution record scored — measure it yourself
| LOOPSruns until doneyes | v0's live docs show a real agent loop: it coordinates multi-step workflows, has Auto-continue mode, opens/tests/critiques apps and fixes proactively, recovering and trying alternatives until done. Step count varies by task. |
| CHOOSESpicks its own toolsyes | Docs: v0 "will automatically use the appropriate capabilities based on what you're asking for" — web search, browser, terminal, MCP picked at runtime per request, not pre-wired. Published NO contradicts its own rationale. |
| ACTSchanges the worldyes | Writes files and runs bash in its sandbox — unit tests, Vercel and GitHub CLI calls. |
| RECOVERSfixes its own errorsyes | Docs confirm rationale: "Recovers from issues and tries alternative approaches" and it diagnoses/fixes errors "automatically as part of the generation loop" while progressing multi-step tasks without approval. Verdict was wrong. |
| UNSUPERVISEDnobody watchingyes | v0 docs name "Full" mode, which skips all terminal permission checks so a full task runs with no per-command approval; "Auto" default also runs allowlisted commands unprompted. Rule 05 satisfied. |
Capability decisions combine documented behaviour and execution evidence. The proof label on each row shows which of the two decided it, and Leash is reported separately, only when it was measured from an execution record. Primary source: v0.app. Tested against: v0 (v0.app docs: agentic features, terminal commands). Our confidence in this dossier: 7/10.
Can you see it work?
the weakest of the five grades aboveDOCS CLAIMED — The vendor's documentation states it and we read the page; we did not watch it happen. Trace access: VENDOR-ONLY — Steps appear in a demo or worked example you cannot run.
Shows agent-action progress, error fixes, version history; no exportable step trace
Where we looked: v0.app/docs and Platform model-api (chats/messages, usage/events endpoints); UI gives 'progress indicators for all agent actions' + error-fix diagnostics but no discrete tool-call/error/timestamp trace · see the evidence →
- 2026-08-114/5 AGENT → 5/5 AGENTRECOVERS corrected to yes: Docs confirm rationale: "Recovers from issues and tries alternative approaches" and it diagnoses/fixes errors "automatically as part of the generation loop" while progressing multi-step tasks without approval. Verdict was wrong.editorial correction — no case was filed
- 2026-08-113/5 WRAPPER WITH AMBITION → 4/5 AGENTCHOOSES corrected to yes: Docs: v0 "will automatically use the appropriate capabilities based on what you're asking for" — web search, browser, terminal, MCP picked at runtime per request, not pre-wired. Published NO contradicts its own rationale.editorial correction — no case was filed
- 2026-08-112/5 WRAPPER WITH AMBITION → 3/5 WRAPPER WITH AMBITIONUNSUPERVISED corrected to yes: v0 docs name "Full" mode, which skips all terminal permission checks so a full task runs with no per-command approval; "Auto" default also runs allowlisted commands unprompted. Rule 05 satisfied.editorial correction — no case was filed
- 2026-08-111/5 CRON JOB WITH EXTRA STEPS → 2/5 WRAPPER WITH AMBITIONLOOPS corrected to yes: v0's live docs show a real agent loop: it coordinates multi-step workflows, has Auto-continue mode, opens/tests/critiques apps and fixes proactively, recovering and trying alternatives until done. Step count varies by task.editorial correction — no case was filed
Badges
earned, never for sale
These render from the live dataset, so a badge cannot outlive the verdict it claims. If the verdict changes the badge changes with it, and an unearned one returns 409 rather than an image. No sponsor can buy one, at any price.
Think this is wrong?
bring evidence, not opinionName the criterion and link a trace, a doc, or a recorded run that shows the behaviour. A verdict changed by evidence is the best thing that can happen to this site, and the change goes on the public record with your reason attached. Vendors are welcome, and a vendor's own filing is labelled as such rather than buried.