wildcard/caro

caro: fast Rust CLI that turns natural‑language tasks into a safe POSIX command. Built for macOS (MLX/Metal) with a built‑in model; supports vLLM/Ollama/LM Studio. JSON‑only output, safety checks, confirmation, multi‑step goals, devcontainer included.

This repository also configures its own agents. See what caro tells them →

40Stars on the repository
158Mods indexed here, across every type
yesterdayLast push, which is what freshness is scored on
AGPL-3.0Licence, which decides whether bodies are shown

wildcard/caro

Command Claude Code

Central dispatcher for all automated development flows. Manages scheduling, execution, and monitoring of Technical, Content, and Management packs.

not rated 40 +1 changed 4d ago A 0 tokens AGPL-3.0

caro-backlog-groom

02

wildcard/caro

Command Claude Code

Scan GH issues / PRs / branches / recent Claude-Code commits and sync them into the beads queue for v1.2.0. Idempotent — safe to run on any cadence.

not rated 40 +1 yesterday A 39 tokens AGPL-3.0

caro-coder-loop

03

wildcard/caro

Command Claude Code

Continuous v1.2.0 feature development — claims next ready beads task, implements via kraken (TDD), opens PR. Loop until no ready work remains.

not rated 40 +1 changed 4d ago A 34 tokens AGPL-3.0

caro.context

04

wildcard/caro

Command Claude Code

Analyze current session context and provide recommendations for maintaining productivity.

not rated 40 +1 yesterday A 0 tokens AGPL-3.0

caro.feature

05

wildcard/caro

Command Claude Code

Unified guided workflow for feature development using Spec-Kitty. Detects current state and suggests next steps.

not rated 40 +1 yesterday A 20 tokens AGPL-3.0

caro.frustrated-qa

06

wildcard/caro

Command Claude Code

Daily 5 AM autonomous QA routine that drives the caro-frustrated-beta sub-agent against the commands the website advertises, classifies the paper-cuts a real power user would hit, scans GitHub for stale work that should have prevented the pain, and turns every finding into a concrete GitHub issue with a fix direction.

not rated 40 +1 yesterday B 0 tokens AGPL-3.0

caro.ml

07

wildcard/caro

Command Claude Code

ML/DS engineering skill for Caro fine-tuning pipeline — dataset growth, base-model selection, training experiments, eval harness extension. Targets M4 Max 48 GB.

not rated 40 +1 yesterday A 35 tokens AGPL-3.0

caro.qa

08

wildcard/caro

Command Claude Code

Quality assurance skill for systematic bug investigation, reproduction, and documentation.

not rated 40 +1 yesterday A 12 tokens AGPL-3.0

wildcard/caro

Command Claude Code

Acceptance-verify a caro release — does it actually deliver what was advertised, what PRs claimed, what beta testers requested? Report gaps as bugs and randomly sample one feature for deep testing.

not rated 40 +1 yesterday D 38 tokens AGPL-3.0

caro.release.hotfix

10

wildcard/caro

Command Claude Code

Emergency hotfix workflow for critical security patches.

not rated 40 +1 yesterday A 8 tokens AGPL-3.0

wildcard/caro

Command Claude Code

Create release branch and run all pre-flight checks for caro release.

not rated 40 +1 yesterday A 12 tokens AGPL-3.0

wildcard/caro

Command Claude Code

Create PR, merge, tag, and publish caro release.

not rated 40 +1 yesterday B 11 tokens AGPL-3.0

wildcard/caro

Command Claude Code

Run security audit and fix vulnerabilities for caro release.

not rated 40 +1 yesterday A 9 tokens AGPL-3.0

caro.release.verify

14

wildcard/caro

Command Claude Code

Verify published caro release on crates.io.

not rated 40 +1 yesterday C 7 tokens AGPL-3.0

wildcard/caro

Command Claude Code

Bump version and update changelog for caro release.

not rated 40 +1 yesterday B 10 tokens AGPL-3.0

caro.roadmap

16

wildcard/caro

Command Claude Code

Project manager skill for roadmap alignment, work selection, and workflow routing.

not rated 40 +1 yesterday B 13 tokens AGPL-3.0

caro.sync

17

wildcard/caro

Command Claude Code

Sync code across different parts of the codebase (roadmap, docs, instructions).

not rated 40 +1 yesterday A 17 tokens AGPL-3.0

discuss

18

wildcard/caro

Command Claude Code

Optional pre-planning discussion phase for complex features. Use when implementation preferences, constraints, or trade-offs need clarification before diving into specifications. Inspired by GSD's discuss phase.

not rated 40 +1 yesterday A 38 tokens AGPL-3.0

idea-sourcing-loop

19

wildcard/caro

Command Claude Code

Automated idea sourcing from relevant sources in our domain (CLI tools, AI, Rust, developer productivity). Extracts actionable ideas for product features, content creation, and marketing opportunities.

not rated 40 +1 changed 4d ago A 0 tokens AGPL-3.0

ml-fine-tune-loop

20

wildcard/caro

Command Claude Code

Autonomous daily routine for the caro-ml-ds-engineer role. Runs at 12 AM PST per .claude/automation/config/schedule.yaml. One pass per 24h, capped.

not rated 40 +1 yesterday B 0 tokens AGPL-3.0

pr-management-loop

21

wildcard/caro

Command Claude Code

Automated PR management that monitors open PRs, interacts with external review agents (Kubic, Copilot), keeps PRs fresh through rebasing, and ensures work flows to completion.

not rated 40 +1 changed 4d ago A 0 tokens AGPL-3.0

prompt-tuner

22

wildcard/caro

Command Claude Code

Improve embedded LLM system prompt based on evaluation test failures.

not rated 40 +1 yesterday A 11 tokens AGPL-3.0

qa-automation-loop

23

wildcard/caro

Command Claude Code

Automated QA testing loop that spawns multiple unbiased beta tester profiles, consolidates feedback, and creates GitHub issues for discovered bugs.

not rated 40 +1 changed 4d ago A 0 tokens AGPL-3.0

wildcard/caro

Command Claude Code

Orchestrate systematic beta testing of model bundles with multiple profiles, collect structured results, and generate QA sign-off report.

not rated 40 +1 yesterday A 28 tokens AGPL-3.0

At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: