Skip to content
CAI
Software that uses CAICheck a score

Michael-A-Kuykendall/shimmy

57.2

Adequate · 14 November 2026

15k

lines of production code

Rust

primary language

5

measurements over time

CAI band scale
CAI trend line
CAI lens gauges

What this system is

Features

Unified Jinja-first prompt rendering and OpenAPI documentation

Prompt construction now uses a single renderer that prioritizes the model's native GGUF Jinja chat template, falling back to hardcoded family templates or raw text, which improves compatibility with diverse model architectures. The API layer has been updated to use this unified rendering for both chat and completion endpoints, and a new \raw\_prompt\ option allows bypassing templates for base models. Additionally, an embedded OpenAPI specification and Swagger UI documentation page are now available to help users explore the API endpoints.

src · high confidence

Behavioural changes

2 commits (0 fixes) modifying Formula

A change to existing behaviour in Formula — 2 commits, 1 file: “shimmy-49m: add versioned licensed Homebrew formula \[skip ci\]”; “shimmy-49m: reject unsupported Mac architecture at install boundary \[skip ci\]”.

Formula · medium confidence · unverified

6 commits (0 fixes) modifying packaging

A change to existing behaviour in packaging — 6 commits, 7 files: “shimmy-tnv: prepare tracker-free Shimmy 2.6.6 source and checksums \[skip ci\]”; “Prepare Shimmy 2.6.5 with exact usage and authenticated deployment”; “shimmy-49m: add versioned licensed Homebrew formula \[skip ci\]”; and 3 more.

packaging · medium confidence · unverified

Airframe GPU engine replaces legacy backends as the default inference path

The engine module has been refactored to use the new Airframe GPU engine as the primary inference backend, replacing the previous HuggingFace, Llama, and MLX engines which have been removed. The \InferenceEngineAdapter\ now instantiates \AirframeEngine\ (when the \airframe\ feature is enabled) and \SafeTensorsEngine\, routing GGUF files to Airframe by default. The \GenOptions\ struct has been extended with new fields (\grammar\_mode\, \fse\_reject\_patterns\, \math\_bypass\, \trace\_path\, \session\_id\, \raw\_prompt\) to support advanced generation controls, and the \ModelSpec\ now includes a \chat\_template\ field to enable Jinja-first prompt rendering, ensuring instruct models are correctly templated before generation.

src/engine · high confidence

Fixes

1 commit (1 fix) fixing src/cache

A fix in src/cache — 1 commit (1 fix), 1 file: “fix: reconcile response cache capacity from stored entries”. Also in this area: Fixes response cache size accounting and eviction statistics.

src/cache · medium confidence · unverified

Consolidated regression test suite and removed obsolete benchmark scripts

The regression testing infrastructure has been consolidated from a large, auto-discovered set of individual test files into three main test targets: core, handlers, and compile\_checks. The automated runner scripts (run-regression-tests.sh and run-regression-tests-auto.sh) have been updated to execute these consolidated suites, reflecting changes such as the addition of SSE regression tests, fixes for template compilation, and updates to error handling and API compatibility checks. Additionally, the standalone Python benchmark and MoE stress test scripts (benchmark.py and moe\_stress\_test.py) have been removed from the repository. Also in this area: API authentication, usage metadata, and prompt rendering fixes. Also in this area: 1 commit (0 fixes) modifying (repo-wide).

scripts · medium confidence · unverified

Test coverage

Consolidated test suite with new compile-time, core, routing, and handler coverage; Regression test suite consolidated and modernized for Airframe engine; Updated PPT contract tests to target Airframe backend.

Dependencies

Shimmy v2.6.4: Airframe GPU engine default and dependency refresh

Shimmy has been updated to version 2.6.4, making the Airframe native WebGPU engine the default inference backend (previously, the default included llama.cpp and HuggingFace features). The \airframe\ dependency is now pinned to version 0.4.3, and the \shimmyjinja\ dependency is added at version 0.5.0. Several underlying dependencies have been updated, including \reqwest\ (0.11 to 0.12), \anyhow\ (1.0.100 to 1.0.104), \async-trait\ (0.1.89 to 0.1.92), and \assert\_cmd\ (2.1.1 to 2.2.2). Legacy llama.cpp feature flags remain as deprecated stubs, and the package description now highlights cross-platform WebGPU acceleration via Airframe.

(dependencies) · high confidence

Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

How this codebase got here

This is the PUBLIC form of this artifact. Findings are listed in full, but the details of SECURITY findings — which rule fired, in which file, on which line, and how to fix it — are deliberately withheld, and any secret-scanner results are excluded entirely. Where detail is absent here it was REMOVED FOR PUBLICATION; it is not missing from the analysis. The complete artifact is available from the repository owner.

Score

  • CAI 60 → 57 (-2.8)
  • Rubric changed (rubric-2026.10.5 → rubric-2026.11.4) — scores are not directly comparable.

Lenses

  • Code Health 55 → 45 (-10.0)
  • Architecture 89 → 89 (+0.1)
  • Maturity 62 → 69 (+6.5)
  • Readiness 56 → 60 (+4.4)
  • Security 74 → 71 (-2.5)
  • Performance 91 → 91 (+0.0)

Resolved (21)

  • Duplicated block (7 lines × 2) (src/api.rs)
  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • Medium CVE: RUSTSEC-2026-0381 (Cargo.lock)
  • Medium CVE: RUSTSEC-2026-0382 (Cargo.lock)
  • Medium IaC: WD-DOCKER-0003 (deploy/Dockerfile)
  • Medium IaC: WD-DOCKER-0003 (deploy/Dockerfile)
  • Medium advisory (unsound): RUSTSEC-2026-0386 (Cargo.lock)
  • Medium vulnerability: [GHSA redacted] (Cargo.lock)
  • Medium vulnerability: [GHSA redacted] (Cargo.lock)
  • No ADRs found
  • Release publish has no approval gate
  • Repeated repair: src/invariant_ppt.rs (src/invariant_ppt.rs)
  • Workflow holding a long-lived secret is unscoped (.github/workflows/release.yml)
  • Workflow token grant is wider than its jobs use (.github/workflows/docker-publish.yml)
  • smoke_text.run (cognitive 33) (scripts/smoke_text.py)
  • …and 1 more

New (42)

  • Duplicated block (10 lines × 3) (src/anthropic_compat.rs)
  • Duplicated block (10–11 lines × 2) (src/anthropic_compat.rs)
  • Duplicated block (17 lines × 2) (src/anthropic_compat.rs)
  • Duplicated block (17 lines × 2) (src/openai_compat/mod.rs)
  • Duplicated block (8 lines × 2) (src/anthropic_compat.rs)
  • Duplicated block (8 lines × 2) (src/metrics.rs)
  • Duplicated block (9 lines × 3) (src/api.rs)
  • Duplicated block (9 lines × 4) (src/anthropic_compat.rs)
  • Duplicated block (9–11 lines × 2) (src/api.rs)
  • Edited copy of a member (11 corresponding lines) (src/prepared_prompt.rs)
  • Edited copy of a member (12 corresponding lines) (src/templates.rs)
  • Edited copy of a member (13 corresponding lines) (templates/frameworks/fastapi/main.py)
  • Edited copy of a member (9 corresponding lines) (src/metrics.rs)
  • FunctionTooLong: shimmy::anthropic_compat::messages (src/anthropic_compat.rs)
  • FunctionTooLong: shimmy::api::generate (src/api.rs)
  • FunctionTooLong: shimmy::openai_compat::completions (src/openai_compat/mod.rs)
  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • High: security finding (details withheld)
  • …and 22 more

Changes since last survey

  • 84 commits — 83 feature/other, 1 fixes

By area

  • (root) — 29 commits
  • .github/workflows — 5 commits
  • Formula/shimmy.rb — 3 commits
  • .github/scripts — 2 commits
  • bucket/shimmy.json — 2 commits
  • docs/FAQ.md — 2 commits
  • docs/FEATURES.md — 2 commits
  • docs/INTEGRATION.md — 2 commits
  • docs/MIGRATION_v2.md — 2 commits
  • docs/quickstart.md — 2 commits
  • .github/instructions — 1 commit
  • deploy/README.md — 1 commit
  • docs/API.md — 1 commit
  • docs/ARCHITECTURE.md — 1 commit
  • docs/BINARY_INSTALL.md — 1 commit
  • docs/CERTIFICATION.md — 1 commit
  • docs/CHAT_TEMPLATES.md — 1 commit
  • docs/CHAT_TEMPLATING.md — 1 commit
  • docs/CONFIGURATION.md — 1 commit
  • docs/CROSS_COMPILATION.md — 1 commit

Notable commits

  • fix: fix: reconcile response cache capacity from stored entries
  • change: Clarify the container model volume in the quickstart
  • change: Document ARM64 download and current engine version
  • change: Document Linux ARM64 binaries and native platform verification
  • change: Keep agent instructions private and document cache correction
  • change: Pin Python 3.12 for container release verification [skip ci]
  • change: Prepare 2.6.7 with qualified Airframe and multilingual release documentation [skip ci]
  • change: Prepare Shimmy 2.6.5 with exact usage and authenticated deployment
  • change: Publish tested native AMD64 and ARM64 images and serialize CI tests
  • change: Resolve published Airframe 0.4.4 for Shimmy 2.6.5
  • change: Simplify SPONSORS.md to remove public individual listing
  • change: Standardize sponsorship language in README (remove public sponsor listing)
  • change: Test four native platforms and publish ARM64 binaries after release gates
  • change: Use the versioned GHCR deployment image and describe actual binary size honestly
  • change: Use the versioned release heading required by local release checks [skip ci]
  • change: chore: minimize sponsor showcase
  • change: chore: minimize sponsor showcase
  • change: docs(i18n): update released Shimmy 2.6.7 Chinese guides
  • change: docs(i18n): update released Shimmy 2.6.7 Chinese guides
  • change: docs(post-release): correct 2.6.7 front-facing RELEASE_GATES_CHECKLIST.md
  • …and 64 more

Architecture

  • Unchanged — 0 containers · 1 contexts · 0 edges

Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

Survey your own repository

Michael-A-Kuykendall/shimmy was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.

About this page

  • The score is its most recent published measurement, taken on 14 November 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
  • Measured at commit 9f6d996ec038a9cc1a41b70e79feee4d7c7e349e — the exact code this score is about.
  • Scored under rubric-2026.11.4 — the same rubric and the same method as every other entry in this index.
  • Measured by Canine Code Analyzer (cca.canine.dev) using cca measure-2026.11.65+preprod-e1a3b69fc8ee.