Repository navigation
v1.10.0
Install
curl -fsSL https://raw.githubusercontent.com/JeremiahM37/grimoire/main/install.sh | sh # Linux, macOS
brew install JeremiahM37/tap/grimoire # Homebrew
scoop bucket add jeremiahm37 https://github.com/JeremiahM37/scoop-bucket && scoop install grimoire # Windows
docker run -p 9111:9111 -v grimoire-vault:/vault ghcr.io/jeremiahm37/grimoire:1.10.0
go install github.com/JeremiahM37/grimoire/go/cmd/grimoire@latest # the module lives in go/, so @latest resolves the go/v1.10.0 tagWindows PowerShell one-liner: irm https://raw.githubusercontent.com/JeremiahM37/grimoire/main/install.ps1 | iex.
Archives (binary beside web/ and plugins/: extract and run), .deb, .rpm and checksums.txt are attached below.
Changelog
- 9697e41 Approvals panel: review and decide agent actions from the web app
- c1c0c1c Benchmarks: utilization-trace validation harness (simulated repo, paired with/without runs, shadow replay)
- 9b27f4d Grounded answers: infer instead of 'not stated' when evidence bears on it; complete lists; best-supported dates
- 4dbaebf Grounded answers: tests for the claude CLI client and answer edge cases
- 891d7d7 Grounded answers: time annotation, entity timeline, gather/answer/check procedure
- 12bfcfd LoCoMo round 11: 92.6% (J) / 91.4% strict overall on the frozen 500; overall badge
- b2a62d5 LoCoMo: multi-hop 89.1% (J) / 84.8% strict with grounded answers and a Sonnet reader; badge
- 315d21f Merge branch 'feat/am-trace-eval' into feat/agent-memory
- 50cf6a4 Merge branch 'feat/answer-grounded' into feat/agent-memory
- 6f95197 Trace: install the doubly-robust estimator so holdout records yield a causal lift (or 'insufficient data'); lint and dead-code cleanup
- 36b58ca agent install --memory-only: the memory hook and outcome tracking without the bank's session-retention hooks
- 7205a70 utilization/estimate: IPS and doubly-robust lift from randomized holdout, EB pooling by kind, exposure-only utility