Pluqqy — Summary

Pluqqy (21 stars) is a Go CLI tool for building and managing "minimal viable context" for AI coding assistants. Users compose reusable components — contexts, prompts, and rules — into named pipelines stored in .pluqqy/. When a pipeline is activated with pluqqy set <name>, it writes a PLUQQY.md file containing all composed instructions. The AI coding tool then references @PLUQQY.md as the active context. Switching instruction sets is as simple as pluqqy set <different-pipeline> — the reference stays the same while the content changes. The tool ships a full TUI (terminal UI) and a comprehensive CLI with JSON/YAML output flags, archived state management, and clipboard integration. Pluqqy is not an orchestration framework or workflow system — it is a context composition manager: the user pre-builds instruction libraries and switches between them for different tasks. Compared to seeds, it occupies a unique niche; no seed directly addresses the problem of managing multiple context/prompt sets for different development contexts, making it closest to agent-os (both use markdown files as the context substrate) but specifically focused on the switch-between-contexts use case rather than a structured methodology.

Pluqqy — Overview

Origin

Created by pluqqy (GitHub: @pluqqy). Go-language implementation. Last commit October 2025.

Philosophy

"Build and manage minimal viable context for your AI coding assistant from your terminal."

"This approach keeps your context minimal and focused - only including what's relevant for the current task. Both you and the AI work from the same single source of truth, eliminating confusion about which instructions are active while preserving context space."

The core insight: developers often have multiple contexts they work within (web dev, CLI dev, data science) and need different instruction sets for each. Copy-pasting prompts is error-prone. Pluqqy provides a named-pipeline system: create once, switch instantly.

Key Concepts

Design Philosophy

• "Claude Code or other AI coding tools, just reference @PLUQQY.md instead of copying and pasting prompts"
• "Need different instructions? Set a different pipeline — the file updates automatically, but you keep referencing the same @PLUQQY.md"

Version

Analyzed from main branch (no semver in README); last pushed 2025-10-03.

Pluqqy — Architecture

Distribution

• Type: CLI tool (Go binary)
• Install: git clone + make install
• Binary name: pluqqy
• Language: Go

Installation

git clone https://github.com/pluqqy/pluqqy-terminal
cd pluqqy-terminal
make install
# Installs to $GOPATH/bin or $HOME/go/bin

Directory Tree

pluqqy/pluqqy-terminal/
├── cmd/
│   ├── pluqqy/         # Main CLI entrypoint
│   └── commands/       # Subcommand implementations
├── pkg/
│   ├── composer/       # Pipeline composition logic
│   ├── examples/       # Example component/pipeline data
│   ├── files/          # File I/O, .pluqqy/ management
│   ├── models/         # Data models (Pipeline, Component, etc.)
│   ├── search/         # Search functionality
│   ├── tags/           # Tag management
│   ├── tui/            # Bubble Tea TUI implementation
│   └── utils/          # Utilities
├── internal/           # Internal packages
├── assets/             # Screenshots
├── Makefile            # Build + install targets
└── go.mod              # Go module definition

# Target project structure (generated by `pluqqy init`):
.pluqqy/
├── pipelines/
├── components/
│   ├── contexts/
│   ├── prompts/
│   └── rules/
├── archive/            # Archived items
│   └── ...
└── tmp/                # Pipeline output files

Required Runtime

• Go (for build from source)
• Any AI coding assistant that can reference @PLUQQY.md

Target AI Tools

Claude Code (primary mention), any AI coding tool that supports file references

Pluqqy — Components

CLI Commands

Pipeline Commands

Management Commands

Example Commands

TUI

Global Flags

--output json|yaml, --quiet, --yes, --no-color

Component Types

• Contexts — .pluqqy/components/contexts/<name>.md — codebase/domain info
• Prompts — .pluqqy/components/prompts/<name>.md — specific instructions
• Rules — .pluqqy/components/rules/<name>.md — behavioral constraints
• Pipelines — .pluqqy/pipelines/<name>.yaml — named composition of components

Pluqqy — Uniqueness

Differs from Seeds

No seed directly addresses Pluqqy's problem. Closest to agent-os — both use markdown files as the content substrate and both are about giving AI agents the right context. The critical difference: agent-os provides a single structured set of standards/profiles and uses slash commands to write them; Pluqqy provides a composable library of components that users switch between for different contexts. Pluqqy is the only framework in this corpus that is explicitly about context management and switching rather than workflow orchestration or methodology. It does not compete with BMAD-METHOD, superpowers, or spec-kit — it could be used alongside any of them (the Pluqqy pipeline could include a BMAD persona as a component).

Positioning

Pluqqy fills the gap between "just use CLAUDE.md" and "adopt a full framework like BMAD". It is a minimalist organization layer: keep your prompts organized, compose them into context-appropriate pipelines, and switch cleanly between development contexts. The CLAUDE.md migration tool makes the value proposition concrete for existing Claude Code users.

Observable Failure Modes

1. Stale PLUQQY.md — If a developer forgets to run pluqqy set before starting work, the AI gets stale context from a previous pipeline.
2. Component drift — As the codebase evolves, components become outdated. There is no automatic staleness detection or audit command.
3. Small ecosystem — 21 stars, last commit October 2025. Risk of abandonment before the tool reaches stable user adoption.
4. No validation — No checks that a pipeline's component references are valid, that components exist, or that the output PLUQQY.md is complete.
5. Concatenation only — Pipeline composition is simple string concatenation. No conditional inclusion, no variable substitution, no inheritance between pipelines.