Skip to main content
World Wide TechnologyBenchAI tool benchmarks
Workflow Tool
StatusWatch
SignalTracked
EvidenceGrade B

Gas Town

48
D 0 vs last quarter

Multi-agent orchestration layer that coordinates 20-30 parallel Claude Code/Codex/Gemini agents via git worktrees, with a Mayor/Polecats/Refinery hierarchy and a Bors-style bisecting merge queue.

UX / DXCapabilityReliabilityValueCommunityEnterpriseAutonomyIntegration

Dimension breakdown

Score · confidence
UX / DX
50% conf45
Capability
50% conf50
ReliabilityIncomplete data at this time
Value
50% conf45
CommunityIncomplete data at this time
Enterprise / Compliance
50% conf20
Autonomy
50% conf70
Integration
50% conf60
Sources blend review platforms, community sentiment, the Signal Radar and practitioner ratings. Scores re-blend each quarter.

Multi-agent orchestration layer that coordinates 20-30 parallel Claude Code/Codex/Gemini agents via git worktrees, with a Mayor/Polecats/Refinery hierarchy and a Bors-style bisecting merge queue. Gas Town v1.0 stable; creator Steve Yegge in maintenance mode with new co-maintainers onboarding while focus shifts to the Gas City SDK rewrite.

Biggest Change

Gas Town by Kilo went GA May 19, 2026 — a fully cloud-hosted version (GitLab co-founder Sid Sijbrandij's startup) on managed infrastructure with 500+ models behind a single Kilo Gateway API and unified billing, removing the heavy self-hosting/tmux ops burden.

This lifts integration and viability.

Caution

enterprise governance (SSO/SCIM/RBAC/audit) lives in Kilo's separate KiloClaw product, NOT Gas Town itself — the tool stays consumer-grade with no SOC 2.

Two trust controversies persist: the $GAS token 98% collapse (Jan 2026) and a newer credit-stealing dispute (issue #3649, Apr 2026) where default installs spent users' LLM credits and GitHub credentials on upstream bug-fixes without disclosure. Steep DX (tmux, Go 1.23+) and ~$100/hr token burn remain.

Tracked; promotion to Assessed gated on hands-on internal testing or a named enterprise customer.

Recommended

Use cases

Not yet assessed — this section fills in as ACES research covers the tool.

Score caps

Risk flags

  • No enterprise features

    enterpriseConditional

    No enterprise customers/features

    Caps all dimensions at 70

    Removed whenEnterprise tier launched with SSO, audit logs, or compliance certification

Assessment

Status rationale

Tracked because Gas Town has well-established analyst coverage (Cloud Native Now, The New Stack, Addy Osmani, Nimbalyst, SE Daily), strong community consensus (35K+ combined stars, 500+ contributors, 2,600+ Discord), a documented v1.0, and now a cloud-hosted GA path via Kilo. Promotion from Tracked to Assessed is blocked by the absence of internal hands-on testing and the lack of a named enterprise customer for Gas Town specifically (Kilo declined to name enterprise customers).

Enterprise blockers persist (no SSO/RBAC/audit/SOC 2 on the orchestrator itself — those are in Kilo's separate KiloClaw product, so the no-enterprise-features cap applies) and two trust controversies ($GAS token collapse; #3649 credit-usage dispute) damage enterprise trust. Assessment remains research-based, not internally validated.

Watch for

Movement triggers

Upgrade to Assessed if: internal hands-on testing completed with documented findings, OR a named enterprise customer running Gas Town (or Gas Town by Kilo) in production is confirmed. Upgrade scores if: Gas Town by Kilo (or Gas City) ships first-class SSO/RBAC/audit on the orchestrator itself (removing the no-enterprise-features cap), formal security audit or SOC 2 completed, the #3649 credit-usage behavior is documented as removed/opt-in, independent benchmark published.

Downgrade if: Kilo cloud effort stalls or is deprecated, additional trust incidents, security breach via unsandboxed agents, community exodus signal, or competitors (Conductor, Antfarm) draw away the at-scale segment.

Caution

Risks & limitations

  • No Enterprise Features

    Moderate

    Radar cap: no-enterprise-features

Capabilities

Integration surface

Not yet assessed — this section fills in as ACES research covers the tool.

Proof points

Adoption & benchmarks

Not yet assessed — this section fills in as ACES research covers the tool.

Spotted something wrong or missing here? Suggest a change →

Per-source contributions

Click any dimension to see the underlying sources and citations.