By adelrioj
Skills for autonomous story execution (dex orchestrator and parallel Codex swarm), adversarial spec review (Codex or local LMStudio), blind multi-model fusion panel (fusion), spec-to-PR conductor (ship-it), OrbStack port-collision migration, and workflow helpers (handoff, sprint status)
Use to raise answer quality on a hard prompt by running it through a blind multi-model panel, then having Opus judge and synthesize one grounded answer. Panelists: 2x Opus subagents, GPT-5.5 via codex, and a local LMStudio model via pi — all over subscription, no API keys. Triggers on: fusion, run through fusion, fuse models, multi-model panel, panel of models, second opinion from multiple models.
Compact the current conversation into a handoff document for another agent to pick up.
Use when a Docker Compose project should stop colliding on host ports across git worktrees or separate projects, by switching containerized services to OrbStack routable domains instead of published host ports. Edits docker-compose and local env files directly, then verifies live. Requires OrbStack as the active Docker engine (aborts otherwise). Triggers on: make orbstack compatible, orbstack migrate, fix docker port collisions, eliminate docker port collisions across worktrees, orbstack ports, switch compose services to orbstack domains, make project orbstack-compatible.
Use when running an already-hardened Superpowers implementation plan through the dex orchestrator (codex backend). Triggers on: convert plan to dex, plan to dex, dex from plan, run plan with dex, plan-to-dex.
Use when a thumbs-upped design spec is ready and you want the whole spec-to-PR pipeline to run autonomously — hardens the spec, writes the plan, executes it via dex, opens a PR, and runs PR review — in one invocation. Triggers on: ship it, ship-it, run the pipeline, spec to PR, autonomous feature pipeline, conductor.
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.
A Claude Code plugin bundling skills for autonomous story execution, adversarial spec review, multi-model answer synthesis, OrbStack port-collision migration, and workflow support.
/plan-to-dex — Run a hardened Superpowers plan through the dex orchestrator. Translates the plan into dex's checkbox-group plan.md (one task = one iteration), imports it, and runs dex apply → dex review autonomously with codex as the fixed backend. Requires the dex and codex CLIs.
/swarm-execute — Implement a feature as parallel user stories: Claude orchestrates via the Workflow (ultracode) tool, Codex workers write all code in isolated worktrees, architect + QA reviews gate every merge. Takes a plain-language request or a plan/spec file directly — no prd.json or tasks/ files. Requires the codex CLI.
/spec-review-codex — Adversarial review of design specs using OpenAI Codex as an independent reviewer. Sends the spec to Codex for rigorous review against a 10-category checklist, fixes CRITICAL and IMPORTANT findings, and loops until the spec passes with zero blocking issues (max 3 iterations). Designed to catch bugs, contradictions, ambiguities, and gaps that self-review misses due to author bias.
/spec-review-local — Same adversarial review loop, but the reviewer is a local model served by LMStudio via the pi CLI. Auto-detects whatever LLM is currently loaded in LMStudio — no hardcoded model. Works fully offline; the reviewer is restricted to read-only tools.
/handoff — Compact the current conversation into a handoff document so a fresh agent can pick up the work. Writes to the OS temp directory, includes a "suggested skills" section, references existing artifacts by path instead of duplicating them, and redacts secrets/PII.
/sprint-status-update — Generate a company-wide sprint recap for Slack from the Notion sprint board: categorizes deliveries, summarizes bug reports, and formats a scannable update. Use on Fridays or at sprint boundaries.
/orbstack-compatible — Make a Docker Compose project stop colliding on host ports across worktrees/projects by moving containerized services onto OrbStack routable domains (<service>.<project>.orb.local). Strips ports: from the base compose, adds an opt-in docker-compose.ports.yml for Windows/CI, rewrites local env URLs to domains, then verifies live. Requires OrbStack as the active Docker engine.
/fusion — Run a prompt through a blind multi-model panel — 2× Opus subagents, GPT-5.5 via codex, and a local LMStudio model via pi — then have a separate Opus judge synthesize the responses into a single verdict. Subscription-only: all panelists use authenticated CLIs or Agent subagents; no API keys required. Pass opus, codex, or local as an optional argument to restrict the panel to a single seat. The local panelist auto-detects whatever model is currently loaded in LMStudio (override base URL with FUSION_LMSTUDIO_URL). Any panelist seat that is unavailable (missing CLI, unloaded model, timeout) is dropped gracefully and the panel continues. Provenance for every run is saved to ~/.claude/fusion-runs/ (never committed). Ported from fusion-fable (MIT); the Gemini/agy seat is intentionally dropped in favor of the local pi/LMStudio seat. Requires: codex CLI authenticated, pi CLI in PATH, LMStudio running with a model loaded (for the local seat).
/ship-it — Run the whole feature pipeline from a thumbs-upped design spec to a reviewed PR in one invocation: hardens the spec with /spec-review-codex, writes the implementation plan, executes it via /plan-to-dex, opens a PR with /commit-commands:commit-push-pr, then runs /review-pr and fixes CRITICAL/IMPORTANT findings (up to 3 passes). A pure conductor — it invokes the existing units live and duplicates none of their logic. The heavy steps (spec-review, plan-writing, dex execution) run as execute-and-report subagents that return only a compact result, keeping their raw codex/dex output out of the conductor's context. Sequential, fully autonomous, best-effort: it never halts on quality, recording any leftover findings in a final report instead. Requires the codex and dex CLIs plus the pr-review-toolkit and commit-commands plugins.
Add the marketplace, then install the plugin:
claude plugin marketplace add adelrioj/claude-skills
claude plugin install claude-skills@claude-skills-marketplace
Restart Claude Code for the plugin to load.
claude plugin marketplace update claude-skills-marketplace
claude plugin update claude-skills@claude-skills-marketplace
claude plugin uninstall claude-skills@claude-skills-marketplace
To also remove the marketplace:
claude plugin marketplace remove claude-skills-marketplace
npx claudepluginhub adelrioj/claude-skills --plugin claude-skillsSpec-driven development using GitHub spec-kit methodology. Constitution-first approach with specify, plan, tasks, and implement phases.
Skills for the Ralph autonomous agent system - Generate PRDs and convert them to prd.json format for autonomous execution
A plugin to help with Ralph's tasks
AI-assisted spec crafting through research, interviews, and multi-LLM review. Like Geppetto carved Pinocchio from wood, transform rough ideas into living implementation plans.
Spec-driven development workflow system with structured phases: Requirements → Design → Tasks → Implementation
Multi-model consensus engine integrating OpenAI Codex CLI, Gemini CLI, and Claude CLI for collaborative code review and problem-solving.