By akashp1712
Manage architectural decisions as a causal graph to reduce context overhead — list, trace, export, and review decisions, rationales, and constraints.
List all active decisions tracked by Crux with their rationales and constraints
Export current session decisions to CLAUDE.md. Use when the user wants to persist architectural decisions beyond the current session.
Show the current Crux decision graph — all active decisions, their rationale, and constraints. Use when asked about architectural decisions, tech choices, or project constraints.
Trace why a specific decision was made by following its causal chain back to the root constraint
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Claude forgets why. Crux doesn't.
Every memory tool remembers what. Only Crux remembers why.
You tell Claude: "Use REST not GraphQL — our team only knows REST."
Forty turns later, Claude suggests GraphQL.
It remembered the decision. It forgot the reason. It ignored the constraint.
Every other memory tool — CLAUDE.md, claude-mem, Cursor memory — stores flat facts. "Use REST."
Claude sees "Use REST" and thinks: that's a suggestion I can override if something better comes along. And it does. Because nothing tells it why REST is non-negotiable.
Crux stores decisions as a causal dependency graph:
CONSTRAINT ⛔ Team only knows REST — GraphQL is not an option
↓
RATIONALE 💡 REST is the only viable choice given team skills
↓
DECISION ▸ Use REST not GraphQL
↑ These are welded together.
Compaction cannot drop one without dropping all.
Claude sees the WHY every single time.
This is the co-inclusion guarantee. The reason travels with the decision. Always.
The pipeline runs on every prompt:
User message
│
▼
┌──────────┐ ┌─────────────┐ ┌──────────┐ ┌─────────┐ ┌────────┐
│ Decompose│──▶│ Deduplicate │──▶│ Select │──▶│ Compile │──▶│ Verify │
│ (extract │ │ (Jaccard │──▶│ (budget │ │ (format │ │(orphan │
│ atoms) │ │ similarity)│ │ + co- │ │ block) │ │ check) │
└──────────┘ └─────────────┘ │ include) │ └─────────┘ └────────┘
└──────────┘
│
▼
additionalContext
injected alongside
the user's prompt
Crux runs as a native plugin for Claude Code.
No npm install. No config files. Works immediately.
# 1. Add the marketplace
/plugin marketplace add akashp1712/claude-marketplace
# 2. Install the plugin
/plugin install crux@akashp1712
(Requires Node.js 18+ and Claude Code)
PreCompact, injects group instructions forbidding Claude from separating a decision from its rationale. On resume, the full graph is reloaded from disk..claude/.crux/sessions/<session_id>.json. Sessions don't interfere. Promote to permanent CLAUDE.md only when you want.local (default): 100% offline, zero-cost, regex-based — extracts CONSTRAINT, RATIONALE, DECISION, GOAL.api: Full semantic extraction via Anthropic or AWS Bedrock. All 12 atom types.hybrid: Local first, falls back to API for substance-dense messages with low local confidence.npx claudepluginhub akashp1712/claude-marketplace --plugin cruxUpstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Comprehensive startup business analysis with market sizing (TAM/SAM/SOM), financial modeling, team planning, and strategic research
v9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.
Permanent coding companion for Claude Code — survives any update. MCP-based terminal pet with ASCII art, stats, reactions, and personality.
Complete creative writing suite with 10 specialized agents covering the full writing process: research gathering, character development, story architecture, world-building, dialogue coaching, editing/review, outlining, content strategy, believability auditing, and prose style/voice analysis. Includes genre-specific guides, templates, and quality checklists.
Comprehensive .NET development skills for modern C#, ASP.NET, MAUI, Blazor, Aspire, EF Core, Native AOT, testing, security, performance optimization, CI/CD, and cloud-native applications