claude

DISCOVER — Find what's worth building

The DISCOVERY front door of the marketplace — upstream of any code. Set a standing discovery /discovery-goal, then run /market-scan: an adversarially-challenged dialogue over a market parameter taxonomy (problem severity, demand evidence, market size, willingness-to-pay, pricing power, competition, reachability, stack-fit) that proposes, scores, validates, and NARROWS candidate opportunities, killing weak ideas early. Loop it over your goal until one candidate earns a keep verdict — the spark. Hands a validated OPPORTUNITY to the ideate plugin when installed, or emits a structured markdown opportunity brief standalone. Self-improving: downstream feedback sharpens the parameters and scoring over time.

0 on GitHubby whatbirdisthatv1.1.1MIT OR Apache-2.0updated 2 months agoworkflow
/

Plugin discover from marketplace idea-to-production in agentic-underground/idea-to-production. Commands read from the repo's manifest — 1 month ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use and it starts local MCP servers. The exact commands are listed below; read them before installing.

What it installs

Skills (4)

  • check

    Verify DISCOVER's external tool dependencies are installed and reachable. Trigger with /discover:check (or "check discover prerequisites"). Runs a fast ✓/✗ probe grouped by tier; advisory by default (never blocks), --strict to fail on a missing required tool. Reads the canonical manifest skills/chec

  • goal-setter

    Set or refine the standing DISCOVERY GOAL that scans run over — the constraints on WHAT to discover (niche, builder edge, price band, stack-fit, effort). Trigger with /discovery-goal (or "set a discovery goal", "set a goal to find a SaaS in <niche>", "refine the goal", "loosen the constraints"). Gua

  • market-scan

    Discover a worth-building opportunity — an adversarially-challenged dialogue that scores and KILLS weak candidates early. Trigger with /market-scan (or "what should I build?", "scan for a niche", "find a market opportunity"). Guard: DISCOVER only; do NOT self-activate during IDEATE/DELIVER — a chose

  • self-improve

    Improve DISCOVER itself — fold downstream feedback into its parameters, scoring, and kill ledger, or self-cleave an over-broad element. Trigger with /discover:self-improve (or "the scanner approved a dud — fix it", "fold this feedback in", "self-improve the market-scan skill"). Applies on a branch →

Slash commands (4)

  • /discover:check

    Verify DISCOVER's external tool dependencies are installed and reachable — a ✓/✗ table by tier (advisory; --strict to fail on a missing required tool).

  • /discover:discovery-goal

    Set or refine the standing DISCOVERY GOAL that scans run over — the constraints on WHAT to discover (niche, builder edge, target price band, stack-fit, effort appetite). Infer-first, one question at a time with a recommended answer + multiple-choice; written to .discover/goal.md.

  • /discover:inspect

    Inspect the DISCOVER plugin itself — audit skills, agents, and knowledge for drift, gaps, and duplication; produce a severity-ranked report.

  • /discover:market-scan

    Discover a worth-building opportunity — an adversarially-challenged ideation dialogue that proposes, scores, and KILLS candidates against a market parameter taxonomy until one earns a keep verdict, then hands a validated opportunity to the ideate plugin (or writes a markdown opportunity brief).

Subagents (2)

  • challenger

    DISCOVER OPPORTUNITY CHALLENGER — the independent skeptic spawned by the market-scan skill before a KEEP opportunity is handed off. A fresh-context second party (NOT the proposing agent) whose sole job is to REFUTE a "keep" verdict: attack market-size assumptions, willingness-to-pay, competitive moa

  • inspector

    DISCOVER INSPECTOR — on-demand agent that audits the DISCOVER plugin it ships in (skills, agents, knowledge, commands under ${CLAUDE_PLUGIN_ROOT}). Triggered by user command only ("inspect DISCOVER" / "/discover:inspect"). Builds a fresh critical-analysis persona, reads every file, and produces DISC

MCP servers (1)

Hooks (1)

  • SessionStart

    bash ${CLAUDE_PLUGIN_ROOT}/hooks/inject-kaizen.sh 2>/dev/null || true

View source on GitHub

More from agentic-underground/idea-to-production

All 8 plugins →

An opinionated, disciplined, quality-first software production system. A conveyor that carries value from IDEA to PRODUCTION across role-tuned agents and value-stations (IDEA ▶ ROADMAP ▶ PLAN ▶ EARS ▶ FEATURE ▶ TEST ▶ IMPLEMENT ▶ STORY ▶ SHIP), governed by three pillars — knowledge parity, quality-first with a performance-delta gate, and waste elimination (muda·mura·muri) — under an overarching token-efficiency constraint. Emits markdown; uses the secure and publish companions automatically when present.

19 skills · 13 commands · 32 agents · 1 MCP server · hooks

0

Turns project artefacts and markdown into publication-grade output: narrative articles mined from git history and docs, standalone diagrams (Graphviz + a Mermaid-native specialist covering the full diagram taxonomy, theming, and ELK), and print-quality PDFs with embedded, A4-legible figures — all held to standard by an adversarial design-reviewer (typography + data-viz canon) that critiques the rendered artefact in a convergent loop. The PUBLISHING companion deliver hands off to when richer-than-markdown output is wanted — or use it directly on any project.

10 skills · 5 commands · 8 agents · hooks

0

The FRONT DOOR, ARRIVAL layer, and meta-layer of the marketplace — the marketplace-level surface the specialist plugins never had. /i2p:help browses the powers you have now, grouped by the DISCOVER ▸ IDEATE ▸ BUILD ▸ DESIGN ▸ SECURE ▸ PUBLISH value flow (listing only what is installed); /i2p:review fans out EVERY installed reviewer — code (deliver), design (design), rendered docs (publish), security (secure) — into ONE adversarial verdict in I2P_REVIEW.md; /i2p:check consolidates every plugin's readiness into one table; /i2p:flow maps where each plugin sits and the next command to run; /i2p:focus declares the active lifecycle phase for context routing so out-of-phase skills stay dormant ('focus, don't uninstall'), broadcast by a SessionStart hook so it survives /clear. It also GREETS whoever opens a repo: a SessionStart hook renders a maintainer-authored .claude/welcome.md (/i2p:define-welcome authors it), and /i2p:statusline turns on the rich two-line idea-to-production status line (context & rate-limit gauges, the product-lifecycle phase, a ⚔ reviewer-catch tally), offered once on first activation. Introduces itself on session start and surfaces occasional ≤25-word tips. A thin orchestrator that delegates to the specialists by capability and never re-implements them. Carries the KAIZEN self-improvement covenant.

10 skills · 10 commands · 1 agent · hooks

0

The OPERATE phase owner — what runs a mission AFTER launch. OPERATE keeps the live product alive and improving: it verifies operational readiness at go-live, instruments and watches the four golden signals (observability), runs incident response (severity, incident-command roles, mitigate-first, runbooks, blameless postmortems), keeps a maintenance cadence (dependency upkeep + CVE patching + tech-debt), and closes the build-measure-learn loop by turning production signals into a new opportunity that re-enters DISCOVER (the cyclic re-entry). /operate-gate is the front door — an operational-readiness check at go-live and the steady-state health view with a READY / WATCH / NOT-READY verdict. Grounded in named canon (SRE SLIs/SLOs/error budgets + golden signals, the Incident Command System, build-measure-learn, ITIL-lite). Stands alone on any live project, or advances the idea-to-production lifecycle OPERATE → DISCOVER (↻) when i2p is installed. Composes the secure plugin's scan-dependencies and hands new opportunities to discover/ideate by capability. Carries the KAIZEN self-improvement covenant: every incident becomes a runbook, every alert gap becomes an SLO.

9 skills · 9 commands · 1 agent · hooks

0

The pre-release security gate for any repository. Catches PII, secrets/credentials, and supply-chain risk before they ship. Runs parallel, multi-agent audits across data files, git history, source, and frontend, and consolidates them into a single severity-ranked report with a PASS / REVIEW / BLOCK verdict. Stands alone, or serves as deliver's SECURE gate when both are installed.

6 skills · 6 commands · 1 agent · hooks

0

The REFINEMENT phase of the marketplace. Takes a validated opportunity (from discover) or a raw idea you already have and refines it to knowledge-parity through an adversarially-challenged dialogue, producing the IDEA PACKAGE: precise AGENT-FACING handoff documents (idea brief, subject-matter-understanding seed, first vertical slice, and a handoff contract that satisfies DELIVER's discovery exit criteria) plus rich, illustrated USER-FACING documents (opportunity narrative, parameter scorecard, market/pricing/competition charts, a user-flow, and a mockup screen) rendered by capability — charts via publish, user-flows/mockups via design (design-reviewed) — when installed; degrades to markdown otherwise. Iterated with the user before hand-off, then carried to PRODUCTION by DELIVER. Supersedes DELIVER's inline ideate, which remains as a graceful-degradation fallback. Also runs /ideate:name — a marketing-grade naming studio: a deep discovery interview (audience, brand archetype, name-type appetite, power-adjacency, humour, risk), wide-net generation across the full name-type taxonomy + affix hooks + phonosemantic tuning, a deterministic token-free checker (registries + neighbour/typo-squat proximity + domain/RDAP + cross-language connotation), and a scored challenge (Neumeier 7 / SMILE-SCRATCH / archetype-fit). Self-improving: ambiguities a downstream builder hits flow back to sharpen the question-bank and the package contract.

4 skills · 4 commands · 2 agents · 1 MCP server · hooks

0

Similar plugins

superpowers

claudecodex

Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques

15 skills · hooks

obra/superpowers 316,836

Backend API design, GraphQL architecture, workflow orchestration with Temporal, and test-driven backend development

9 skills · 1 command · 8 agents

wshobson/agents 42,461

beads

claudecodex

AI-supervised issue tracker for coding workflows

1 skill · 1 agent · hooks

gastownhall/beads 28,698

MiniMax AI skills library for frontend, fullstack, Android, iOS, shader, GIF sticker, document, presentation, spreadsheet, and multimodal media workflows

18 skills

MiniMax-AI/skills 13,828

worktrunk

claudecodex

Worktrunk is a CLI for Git worktree management, designed for parallel AI agent workflows. This plugin provides configuration guidance (LLM commit messages, project hooks, worktree paths), automatic activity tracking (🤖/💬 indicators in `wt list` showing active Claude sessions), and a `/wt-diff` pane showing each worktree's changes.

2 skills · hooks

max-sixty/worktrunk 12,481

gsd-core

claude

GSD Core is a meta-prompting, context engineering, and spec-driven development system for AI coding agents.

72 skills · 72 commands · 64 agents · hooks

open-gsd/gsd-core 11,494

Maintain DISCOVER — Find what's worth building?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins8 plugins