RepoRadar
Leave a Tip
Item detail
github.com

ag2ai/ag2

RepoRadar surfaced ag2ai/ag2 — a developer tool — into the AI Agent Framework section, where it sits at Gold tier with a 'try now' verdict. Its strongest signal is workflow potential, scored 10.0 out of 10.

Open on GitHub →Back to radar
Score9.3
Popularity98.0
Riskconditional
TierGold
Score breakdown
Usefulness9.0
Novelty8.0
Momentum9.0
Maturity9.2
Open-source/build8.4
Evidence7.2
Workflow potential10.0
Setup ease4.2

Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.

Why it matters

Use it when you need coordinated AI agents with stronger conversation-state handling than ad-hoc chains.

Who should use it

developers building production agent platforms builder teams moving from single-agent scripts to workflows platform teams standardizing agent orchestration

Who should skip it

Skip ag2ai/ag2 if the source repository or demo is inactive, unmaintained, or no longer matches the description shown here.

About this signal

ag2ai/ag2 is tracked by RepoRadar as a developer tool in the AI Agent Framework section. First seen 2026-06-18; the source record was last checked on 2026-06-18. The current verdict is 'try now' with a Gold tier and advanced setup difficulty. ag2ai/ag2 leads on workflow potential (10.0) and maturity (9.2); its lowest signal is setup ease (4.2), so factor that in before investing setup time. This page summarizes the public evidence on the linked source page and states where additional review is still needed.

How this item is evaluated

The ag2ai/ag2 record combines a 9.3/10 composite score with separate popularity (98.0), risk (conditional), and setup (advanced) signals. See the scoring methodology for the current weights and evidence definitions.

Putting this into practice? Read How to vet an AI agent or MCP server before you wire it in for the checklist behind this score.

Risk explanation

Agent runtimes can execute shell or API actions through connectors; scope credentials and tools tightly; Concurrency and retry behavior can produce duplicate external actions without strict idempotency.

Evidence links
Closest alternatives / related signals
agents agent-framework automation orchestration a2a
Verification record

What RepoRadar actually verified

Tested in a bounded workflow

Bounded representative workflow retained by RepoRadar verification harness. Last checked 2026-07-14T05:43:53.522446Z.

passed · cohort-20260712-ag2-local-agent

Tester
RepoRadar automated local verification harness
Started
2026-07-13T03:21:53.757898Z
Completed
2026-07-13T03:21:56.019599Z
Environment
Windows 10 AMD64; Python 3.11.9; credential-stripped child environment; disposable home/cache
Install/setup time
0 minute(s)
Evidence scope
Bounded setup or capability check
Cleanup
Per-check temporary home and work directory removed. Shared cohort package cache removed.
Actions exercised
  • Created a disposable home, work directory, and isolated package cache with credential-like environment variables excluded.
  • Executed bounded check: Construct an AG2 conversational agent with LLM use disabled.
  • Captured the complete sanitized stdout, stderr, exit status, and 2.26-second wall time.
Observed results
  • Command exited 0 after 2.26 seconds.
  • Expected marker 'CHECK_OK' was observed in retained output.
Observed strengths
  • AG2 constructed a named ConversableAgent with LLM use and human input disabled, retaining the expected smoke name.
Friction
  • The command emitted stderr; warnings or errors are preserved in the retained log for review.
Limitations
  • No inter-agent conversation, code executor, remote model, or provider integration was exercised.
  • This bounded cohort check is not a production benchmark or a claim of real user-workflow adoption.

Pricing assessment: No paid plan or metered provider usage was exercised; package or licensing, hosting, and provider costs remain workflow-dependent.

Privacy assessment: No repository content, user data, or provider prompt was transmitted; broader product data handling was not assessed by this bounded run.

Open retained test log →

passed · cohort-20260712-ag2-deterministic-agent-routing

Tester
RepoRadar automated local verification harness
Started
2026-07-13T10:35:45.012971Z
Completed
2026-07-13T10:35:47.508214Z
Environment
Windows 10 AMD64; Python 3.11.9; credential-stripped child environment; disposable home/cache
Install/setup time
1 minute(s)
Evidence scope
Bounded representative workflow
Cleanup
Per-check temporary home and work directory removed. Shared cohort package cache removed.
Actions exercised
  • Created a disposable home, work directory, and isolated package cache with credential-like environment variables excluded.
  • Created 1 synthetic fixture file(s) inside the disposable work directory; retained hashes prove the exact inputs.
  • Created two AG2 ConversableAgent instances with LLM and human input disabled, then registered a deterministic reply handler on the worker.
  • Initiated a one-turn outage-routing chat and asserted the two-message history and ROUTE page summary before exporting proof.
  • Executed bounded check: Run a deterministic two-agent AG2 conversation that routes a synthetic outage to on-call.
  • Captured the complete sanitized stdout, stderr, exit status, artifact checks, and 2.49-second wall time.
Observed results
  • Command exited 0 after 2.49 seconds.
  • AG2 delivered the outage message to the worker, invoked the registered reply path, and returned a two-message chat with the expected page route.
  • Expected marker 'CHECK_OK messages=2 summary=ROUTE_page' was observed in retained output.
  • Validated result.json: 3 required marker(s) present and 0 excluded marker(s) absent; size and SHA-256 are retained.
Observed strengths
  • AG2's conversation lifecycle and reply registration supported a complete local inter-agent exchange without provider configuration.
Friction
  • AG2 emits a maximum-turn termination notice even for the intentional one-turn bounded exchange; the notice is retained in the log.
  • Setup or runtime emitted 3 stderr line(s); the complete warnings/errors are preserved in the retained log.
Limitations
  • The registered local reply validates agent messaging and chat history, not LLM-backed reasoning, code execution, group chat, persistence, or remote tools.
  • This credential-free disposable workflow does not establish production scale, model quality, reliability under sustained use, or team adoption.

Pricing assessment: The two-agent exchange used local deterministic reply logic and the open-source AG2 package without a model or hosted runtime.

Privacy assessment: The synthetic outage message stayed inside two in-process agents; no provider, telemetry target, code executor, or external transport was configured.

Open retained test log →

passed · cohort-20260714-ag2-deterministic-agent-routing

Tester
codex-local-audit
Started
2026-07-14T05:43:50.735215Z
Completed
2026-07-14T05:43:53.522446Z
Environment
Windows 10 AMD64; Python 3.11.9; credential-stripped child environment; disposable home/cache
Install/setup time
1 minute(s)
Evidence scope
Bounded representative workflow
Cleanup
Per-check temporary home and work directory removed. Shared cohort package cache removed.
Actions exercised
  • Created a disposable home, work directory, and isolated package cache with credential-like environment variables excluded.
  • Created 1 synthetic fixture file(s) inside the disposable work directory; retained hashes prove the exact inputs.
  • Created two AG2 ConversableAgent instances with LLM and human input disabled, then registered a deterministic reply handler on the worker.
  • Initiated a one-turn outage-routing chat and asserted the two-message history and ROUTE page summary before exporting proof.
  • Executed bounded check: Run a deterministic two-agent AG2 conversation that routes a synthetic outage to on-call.
  • Captured the complete sanitized stdout, stderr, exit status, artifact checks, and 2.79-second wall time.
Observed results
  • Command exited 0 after 2.79 seconds.
  • AG2 delivered the outage message to the worker, invoked the registered reply path, and returned a two-message chat with the expected page route.
  • Expected marker 'CHECK_OK messages=2 summary=ROUTE_page' was observed in retained output.
  • Validated result.json: 3 required marker(s) present and 0 excluded marker(s) absent; size and SHA-256 are retained.
Observed strengths
  • AG2's conversation lifecycle and reply registration supported a complete local inter-agent exchange without provider configuration.
Friction
  • AG2 emits a maximum-turn termination notice even for the intentional one-turn bounded exchange; the notice is retained in the log.
  • Setup or runtime emitted 3 stderr line(s); the complete warnings/errors are preserved in the retained log.
Limitations
  • The registered local reply validates agent messaging and chat history, not LLM-backed reasoning, code execution, group chat, persistence, or remote tools.
  • This credential-free disposable workflow does not establish operator use, production scale, model quality, reliability under sustained use, or team adoption.

Pricing assessment: The two-agent exchange used local deterministic reply logic and the open-source AG2 package without a model or hosted runtime.

Privacy assessment: The synthetic outage message stayed inside two in-process agents; no provider, telemetry target, code executor, or external transport was configured.

Open retained test log →

Verification sources

Longitudinal intelligence

How this decision record is moving

Raw history JSON →

47 dated snapshots retained from 2026-06-18 through 2026-08-13; see the snapshot index for explicit coverage gaps. Stars, version, release, pricing, integration, risk, maintenance, verdict, score, and momentum fields remain explicit even when a source has not reported them. Repository momentum is a normalized 0–10 RepoRadar signal; GitHub stars appear only where the popularity monitor retained exact timestamped observations.

RepoRadar score9.3 current · +0.0 net
Repository momentum9.0 current · +0.0 net
GitHub stars (observed)4,855 current · +90 net
GitHub stars4,855 exact observation
Versionv1.0.1
Last release2026-07-29T05:09:51Z
Maintenanceactive
Current riskconditional
Current verdicttry now
Pricing baselineNo structured commercial pricing baseline
Pricing checkedNot applicable or not recorded
Pricing freshnessNo dated commercial pricing review
Integrations baselineModel Context Protocol

Recent dated points

DateScoreMomentumStarsRiskVerdictMaintenance
2026-08-139.39.04,855conditionaltry nowactive
2026-08-129.39.34,854conditionaltry nowactive
2026-08-119.39.34,851conditionaltry nowactive
2026-08-109.39.34,847conditionaltry nowactive
2026-08-099.39.34,846conditionaltry nowactive
2026-08-089.39.34,840conditionaltry nowactive
2026-08-079.39.34,829conditionaltry nowactive
2026-08-069.39.0Not recordedconditionaltry nownot recorded
2026-08-059.39.0Not recordedconditionaltry nownot recorded
2026-08-049.39.34,829conditionaltry nowactive
2026-08-039.39.34,829conditionaltry nowactive
2026-08-029.39.34,825conditionaltry nowactive

Why the record changed

stars changed

Stars changed: 4854 → 4855.

verification changed

Verification changed: Re-tested recently → Tested in a bounded workflow.

stars changed

Stars changed: 4851 → 4854.

stars changed

Stars changed: 4847 → 4851.

stars changed

Stars changed: 4846 → 4847.

stars changed

Stars changed: 4840 → 4846.

stars changed

Stars changed: 4829 → 4840.

stars changed

Stars changed: 4825 → 4829.

stars changed

Stars changed: 4822 → 4825.

stars changed

Stars changed: 4807 → 4812.

version changed

Version changed: v1.0.0 → v1.0.1.

stars changed

Source-observed stars changed: 4805 → 4807. This reports the retained observation delta and does not infer why the upstream change occurred.