awesome-repositories.com
Blog
MCP
awesome-repositories.com

Discover the best open-source repositories with AI-powered search.

ExploreCurated searchesOpen-source alternativesSelf-hosted softwareBlogSitemap
ProjectMCP serverAboutHow we rankPress
LegalPrivacyTerms
© 2026 Bringes Technology SRL·VAT RO45896025·hello@awesome-repositories.com
michael-denyer avatar

michael-denyer/pstack-claude

0
View on GitHub↗

Pstack Claude

pstack-claude is a plugin and skills pack for AI coding assistants. It bundles rigorous engineering workflows, step-by-step playbooks, routing hooks, model settings and subagent definitions, and translates them so the same content runs natively on many coding harnesses. Its skills are derived from upstream sources, with local forks and translations preserved and licensing provenance recorded.

The distinctive parts are cross-harness porting and multi-agent orchestration. Tool names, model names and subagent types are pretranslated into each runtime's equivalents before installation, so nothing is translated at runtime. Native installs are provided for Claude Code, Codex, GitHub Copilot, OpenCode, Kimi, Antigravity, Pi, Grok Build, OMP, Muse and MiniMax Code, among others. Work can be delegated to background specialist agents, each concurrent writer in its own git worktree. Parallel attempts can be fanned out and combined into one report. A review panel of models from different vendors stress-tests diffs, and a per-role override sheet sets the model and reasoning effort for each role.

Playbooks cover planning, features, test-first and evidence-backed bug fixes, refactoring, performance, investigations, prototypes and long multi-day efforts. Pull request skills watch review and CI, fix failures and merge conflicts, triage feedback, and prepare changes for merge. Other skills assess a change's blast radius, write and strip AI-sounding wording from docs, and build project-local formal verification skills that use TLA+ models and statecharts. A session-start hook routes non-trivial tasks into the workflow, and a setup skill toggles routing and configures models.

Consistency is enforced by generating shared facts from a single source into every committed file, with a pre-commit and CI check that fails on drift. Further checks validate skill definitions, model sheets, plans and cross-harness behavior.

Features

  • Cross-Harness Workflow Translation - Translates skills, tools, models, and subagent types from one coding tool into native installs for many others.
  • Coding Agent Plugin Packages - Bundles skills, routing hooks, model settings and subagent definitions as plugins installable across Claude Code, Codex, Copilot and other coding assistants.
  • Agent Engineering Playbooks - Guides coding agents through planning, test-first bug fixes, refactoring, performance work and investigations with step-by-step playbooks and evidence-backed results.
  • Agent Tool Name Mappings - Maps tool, model, and subagent names from one coding harness to equivalents on other runtimes so the same workflows run unchanged.

AI search

Explore more awesome repositories

Describe what you need in plain English — the AI ranks thousands of curated open-source projects by relevance.

Start searching with AI
4,131 stars·374 forks·JavaScript·MIT·1 view
  • Skill Installation - Converts tool-specific primitive names into the target runtime's equivalents before install, avoiding token cost and mapping errors at runtime.
  • Multi-Platform Skill Runtimes - Ships one shared tree of agent skills to many coding runtimes, each with its own manifest and tool-name mapping.
  • Background Agent Runs - Dispatches child agents into background sessions, each concurrent writer in its own git worktree, with messaging, listing, and stopping.
  • Multi-Model Diff Reviews - Runs a panel of different models that try to break a diff, surfacing bugs, security issues and maintainability problems before merge.
  • Cross-Harness Workflow Porting - Rewrites tool names, subagent types, and orchestration notes through a documented substitution list so workflows built for one harness run in another.
  • Engineering Task Playbooks - Guides planning, features, refactoring, performance, investigations, prototypes, PR upkeep and long projects with ready-made step-by-step playbooks.
  • Multi-Agent Skill Installers - Installs one shared tree of skills, playbooks and subagent roles as native plugins across many AI coding assistants.
  • Parallel Agent Runs - Creates a separate git worktree for each concurrent writing agent so parallel changes never collide, with fallback to separate agent calls.
  • Ready-Made Agent Workflow Skills - Provides a set of prebuilt agent workflow skills for rigorous coding work that run across many coding harnesses.
  • Session Start Task Routing - Routes each incoming task to the matching workflow mode at session start using hooks or injected instructions across supported tools.
  • Step-by-Step Task Plans - Provides ready-made step-by-step playbooks for planning, features, refactoring, performance, investigations, prototypes, pull request upkeep, and long projects.
  • Forensic Investigation Playbooks - Matches each task to a playbook such as bug fix, feature, refactoring, or investigation and follows its steps as written.
  • Engineering Workflow Playbooks - Ships a library of step-by-step playbooks covering planning, features, bug fixes, refactoring, performance, investigations and pull request upkeep.
  • Parallel Subagent Orchestrators - Delegates work to background specialist agents, each concurrent writer in its own git worktree, and fans out parallel attempts into one combined report.
  • Agent Delegation - Delegates design, comparison, and planning roles to specialist agents before implementation and reviews their results with tests and measurements.
  • Skill Directory Configurations - Exposes each skill as a directory with portable name and description metadata so compatible agents discover skills automatically without per-runtime rewrites.
  • Agent Skill Installers - Bundles agent skills and agents into a plugin directory with a manifest and marketplace file so the whole set installs through one plugin command.
  • Skill Routing Indexes - Routes each situation to a matching skill, such as design exploration, adversarial review, or parallel fan-out, before acting.
  • Worktree Isolation - Gives each concurrent writing agent its own git worktree so parallel changes never collide.
  • Skill Primitive Translations - Converts tool-specific skill names and primitives into Claude Code equivalents before installation, so agents avoid runtime translation and mapping mistakes.
  • Long Autonomous Runs - Carries long, multi-day efforts to completion, coordinating many stacked pull requests and subagents without stopping.
  • Role-Based Model Assignment - Detects available models, confirms a choice for each role and review-panel slot, and saves an override sheet that later sessions honor.
  • Model Role Mappings - Maps each agent role and review-panel slot to a model and reasoning effort in one persisted, validated sheet.
  • Multi-Vendor Model Panels - Runs review roles on a panel of several different models by default, defined as a single consistent fact.
  • Parallel Sub-Agent Workers - Spawns several background worker agents over slices or competing attempts, each isolated in its own worktree, and merges their output into one report.
  • Reasoning Effort Selection - Lets a model-override sheet pin a reasoning effort per agent role so subagent dispatches run at the chosen level.
  • Session-Start Routing Hooks - Fires a lightweight hook on startup, resume, clear, and compact to load routing instructions, with fallbacks where hooks are unsupported.
  • Subagent Document Range Delegation - Spawns child agents from a description and prompt, runs them in the background and resumes them by session ID.
  • Subagent Orchestrations - Dispatches subagents by role, mapping general work to coder, read-only work to explore and planning to plan agents.
  • Complexity-Based Task Routing - Loads a short routing instruction at session start and invokes the full workflow only when a task meets complexity conditions.
  • Pull Request Automation - Ships ready-made skills that fix failing CI, resolve merge conflicts, fetch pull request comments and make pull requests easier to review.
  • Generated Copy Consistency Checks - Stamps shared facts from one source into every committed file and fails pre-commit and CI checks when any copy is stale.
  • Generated File Freshness Checks - Runs a pre-commit and CI check that fails when generated files or plugin invariants are stale, catching drift locally before merge.
  • Merge-Ready PR Playbooks - Runs playbooks that watch a PR or stack, triage review-bot findings, verify each PR independently, and stop at merge-ready.
  • Pull Request Babysitters - Watches pull requests through review and CI, fixing failures and feedback until the change is merge-ready.
  • Pull Request Shepherding - Drives pull requests or stacks through conflicts, review threads and CI checks, independently verifies them and lands the verified run.
  • Agentic Task Workflow Definition - Routes multi-file, design-heavy or unknown-cause tasks through a rigorous multi-step agent workflow before any work begins.
  • Plugin Manifests - Ships a separate plugin manifest and tool-name mapping for each coding runtime, applied before installation.
  • Engineering Principles - Grounds each decision in named principles such as minimal change, boundary discipline, and root-cause fixing, citing the principle behind every choice.
  • Human-Reviewed Upstream Syncs - Syncs upstream skill content into translated copies while keeping local translations and declared forks, advancing a pinned revision only when checks pass.
  • Automated Fix Verifiers - Reproduces a failure, delegates the fix, and reruns the failing case to return the change with before-and-after evidence.
  • Agentic Code Reviews - Applies multi-pass review skills and dedicated review subagents to inspect code changes and surface quality problems in depth.
  • Evidence-Backed Bug Fix Workflows - Guides agents to reproduce a failure, trace its cause, delegate the fix, and rerun the failing case with before-and-after evidence.
  • Project-Local Verification Skills - Builds a project-local skill that formally verifies one risky property with a TLA+ model and draws the model as a statechart.
  • Formal Verification Tools - Builds a project-local skill that formally verifies the riskiest untested property, shows it failing on a broken variant, and stays rerunnable.

    • Star history

    Star history chart for michael-denyer/pstack-claudeStar history chart for michael-denyer/pstack-claude

    How this analysis was created: This summary and feature list are AI-generated from collected project material and can contain mistakes. Stars, license and language are imported from GitHub. Inclusion does not mean that we have tested or audited this project. Check the source documentation for any feature you depend on. Learn more on our About page.

    Frequently asked questions

    What does michael-denyer/pstack-claude do?

    pstack-claude is a plugin and skills pack for AI coding assistants. It bundles rigorous engineering workflows, step-by-step playbooks, routing hooks, model settings and subagent definitions, and translates them so the same content runs natively on many coding harnesses. Its skills are derived from upstream sources, with local forks and translations preserved and licensing provenance recorded.

    What are the main features of michael-denyer/pstack-claude?

    The main features of michael-denyer/pstack-claude are: Cross-Harness Workflow Translation, Coding Agent Plugin Packages, Agent Engineering Playbooks, Agent Tool Name Mappings, Skill Installation, Multi-Platform Skill Runtimes, Background Agent Runs, Multi-Model Diff Reviews.

    Which projects share features with michael-denyer/pstack-claude?

    Projects with overlapping indexed features include: rlaope/oh-my-hermes — Oh My Hermes (Hermes Orchestrator) is a workflow layer that sits on top of the Hermes agent runtime. It routes… vinvcn/mattpocock-skills-zh-cn — This repository is a Simplified Chinese localization of a coding-agent skills collection. The skills are plain… gamedev-skills/awesome-gamedev-agent-skills — This repository is a curated collection of portable agent skills that teach AI coding assistants how to build games.… can1357/oh-my-pi — oh-my-pi is an agentic workflow automation platform and AI coding agent orchestrator designed for autonomous software… wheresryan22/anatomy — Anatomy is an agent skill that generates detailed, interactive isometric illustrations of real machines and complex… superdesigndev/treg — Treg is a gateway that lets AI agents search a catalog of third-party API endpoints and call any of them with a single…

    Projects sharing features with Pstack Claude

    These projects share indexed features with Pstack Claude. Shared tags can include platform or build tooling; verify the primary use case before treating a result as a replacement.
    • rlaope/oh-my-hermesrlaope avatar

      rlaope/oh-my-hermes

      5,229View on GitHub↗

      Oh My Hermes (Hermes Orchestrator) is a workflow layer that sits on top of the Hermes agent runtime. It routes requests to staged skills, external coding executors, and verification gates, through chat or a terminal interface. Completion is evidence-gated: every status carries a stage and confidence, and work is marked verified only after tests, reviews, or CI checks have actually passed. It is also distributed as a catalog of installable workflow skills and plugin hooks, for Hermes and other coding-agent hosts. Its main focus is multi-agent coding delegation. Scoped tasks are handed to codin

      Pythonagentaiai-agent
      View on GitHub↗5,229
    • vinvcn/mattpocock-skills-zh-cnvinvcn avatar

      vinvcn/mattpocock-skills-zh-CN

      5,600View on GitHub↗

      This repository is a Simplified Chinese localization of a coding-agent skills collection. The skills are plain Markdown instruction documents that guide coding agents through repeatable engineering tasks: planning, test-driven development, debugging, code review, and issue triage. Commands, code, and file paths stay unchanged in the translation, which tracks the English upstream through periodic merges, each checked by a person before it is accepted. The skills cover several stages of the development workflow. Planning and alignment skills interview the user until a plan or design is fully re

      JavaScriptai-agentchinesedeveloper-skills
      View on GitHub↗5,600
    • gamedev-skills/awesome-gamedev-agent-skillsgamedev-skills avatar

      gamedev-skills/awesome-gamedev-agent-skills

      7,596View on GitHub↗

      This repository is a curated collection of portable agent skills that teach AI coding assistants how to build games. Each skill is a self-contained folder in the SKILL.md format, so it can be loaded by Claude Code, Cursor, Codex, Copilot, Gemini CLI, Kiro and other compliant agents. Engine-specific skills cover Godot, Unity, Unreal, Roblox, and browser or code-first frameworks, with guidance pinned to a stated engine version. A routing layer sets the collection apart. It scans a project's build files to detect the engine and version, classifies the request by task type, and loads only the mat

      Pythonagent-skillsai-agentsantigravity
      View on GitHub↗7,596
    • can1357/oh-my-pican1357 avatar

      can1357/oh-my-pi

      38,142View on GitHub↗

      oh-my-pi is an agentic workflow automation platform and AI coding agent orchestrator designed for autonomous software engineering. It functions as a multi-model LLM router and an LSP-integrated development environment, coordinating specialized AI agents to perform codebase analysis, automated refactoring, and complex task execution. The system distinguishes itself through the use of subagent coordination to execute parallel tasks within isolated environments and an auto-research framework for iterative experiments. It employs AST-driven structural search for code discovery and content-hash an

      TypeScriptai-agentai-coding-agentanthropic
      View on GitHub↗38,142
    Compare all 30 related projects→