v2.8.0 — Review Evidence That Holds Up When Work Gets Real
Gentle AI v2.8.0 strengthens the boundaries that matter when work crosses agents, worktrees, platforms, and recovery paths: review authority remains explicit, SDD recovery remains truthful, runtime state is easier to inspect, and release delivery remains verifiable end to end.
Provenance
- Previous stable:
v2.7.0→4e4470b3 - Release target:
v2.8.0→890ca948e3430abe8d586ed28773a1c7c9c7af65 - Release workflow: run 34753662597
- Verification: preflight, release, and remote-asset verification jobs all succeeded.
- Provider contract:
1.2.0
What's new
Review and Receipt-Driven Development are more explicit at the edges
v2.8.0 hardens review lifecycle behavior where authority, candidate identity, remediation, and terminal outcomes meet. The result is less implicit state, clearer recovery behavior, and tighter protection against stale or misrouted review actions.
Included work: #4383, #4384, #4418, #4438, #4445, #4458, #4472, #4486, #4514, and #4515.
SDD authority and recovery paths are more truthful
SDD execution now gives more precise authority, recovery, and continuation behavior. Failed, interrupted, and resumed work is handled with stricter ownership and clearer state instead of relying on ambiguous progress signals.
Included work: #3023, #4420, #4442, #4454, #4456, #4464, #4500, #4505, #4518, #4519, #4522, and #4531.
Runtime telemetry is easier to inspect and trust
The runtime and dashboard work in this release improves visibility into active operations and their outcomes, including subagent attribution across supported runtimes. Diagnostics become more actionable without turning telemetry into an independent source of authority.
Included work: #4358, #4359, #4362, #4368, #4381, #4447, #4448, #4449, #4450, #4478, and #4483.
Breaking changes
No breaking changes are declared for v2.8.0. The provider contract remains at 1.2.0 and previously published contract artifacts remain byte-frozen.
Upgrade now
brew upgrade gentle-ai
gentle-ai syncRun gentle-ai sync after upgrading so managed agent, reviewer, and runtime assets match the installed binary.
What was fixed
Platform, installation, and uninstall
Installation, upgrade, platform detection, and uninstall behavior received targeted hardening across supported environments.
Included work: #2035, #4001, #4092, #4240, #4351, #4404, #4476, #4485, and #4535.
Remote and release safety
Release and remote-operation paths now have additional safety controls around explicit authorization, trusted inputs, source identity, and publication behavior.
Included work: #4327, #4344, and #4352.
Documentation
Documentation now better reflects the current operational model and supported workflows.
Included work: #4328, #4349, and #4511.
Numbers
117 non-merge commits, 48 merged pull requests, and 11 unique GitHub PR authors since v2.7.0.
Install
brew install Gentleman-Programming/tap/gentle-aior
go install github.com/gentleman-programming/gentle-ai/v2/cmd/gentle-ai@v2.8.0Windows remains supported through the Go installation path. Official Windows binary archives and Scoop publication remain unavailable; Windows upgrades fail closed to Go-based guidance rather than downloading unsigned binaries.
Signed platform archives below: linux_amd64, linux_arm64, darwin_amd64, darwin_arm64, plus gentle-ai-review-provider-contract-1.2.0.tar.gz and gentle-ai-release-provenance-v1.tar.gz. checksums.txt is signed with the release Minisign key (checksums.txt.minisig); see docs/release-signing.md for verification.
Full changelog: v2.7.0...v2.8.0