Claude Night Market

Claude Code plugins for software engineering workflows.

This repository adds 16 plugins to Claude Code for git operations, code review, spec-driven development, and issue management. Each plugin functions independently to allow selective installation while sharing a common testing framework for consistency.

Features

Governance & Quality Hooks adapt context based on the active agent. pensive tracks usage frequency and failure rates to identify unstable workflows. imbue enforces test-driven development via a PreToolUse hook that verifies the existence of test files before allowing implementation writes (markdown modules in modules/ and commands/ directories are exempt since they are validated by abstract:skills-eval). For complex tasks, imbue:rigorous-reasoning requires step-by-step logic checks before tool execution.

Security & Session Management leyline manages OAuth flows for GitHub, GitLab, and AWS with local token caching. conserve implements permission checks, automatically approving safe commands like ls while blocking high-risk operations like rm -rf /. sanctum isolates named sessions for debugging, feature work, and PR reviews. Quality gates in /create-skill and /create-command halt execution if the project has failing tests.

Maintenance & Resilience /cleanup orchestrates bloat removal, code refinement, and AI hygiene auditing in a single pass with progressive depth levels. /update-ci reconciles pre-commit hooks and GitHub Actions workflows with recent code changes, detecting renamed files and orphaned references. recommends updates based on plugin stability metrics and includes module auditing to detect orphaned or missing skill references. attempts to repair failed runs by analyzing previous errors. generates or validates Makefile targets for all documented commands across plugins, with automatic language detection for Python, Rust, and TypeScript projects. 's homeostatic monitor tracks skill stability after each invocation and auto-triggers the agent when degradation is detected, with rollback review via GitHub issues. For strategic decisions, uses a Type 1/2 reversibility framework to route choices to appropriate expert subagents, with enabling embedded escalation at decision points during implementation. adds for agent-level crash recovery and context overflow handling, and for 4-tier task risk gating (GREEN/YELLOW use heuristic matching, RED/CRITICAL escalate to war-room-checkpoint). Agents that modify files can run in worktree isolation () to prevent filesystem conflicts during parallel execution; read-only audit agents support background execution () for non-blocking analysis.