Intelligentium is a curated marketplace of skills, agents, hooks, MCP servers, and tools — versioned, deterministic, and installable in one command via APM.
Every plugin follows the same predictable shape — so your agents can install, compose, and trust them without surprises.
Add the marketplace once, then pull any plugin with apm install <plugin>@intelligentium — no manual wiring.
Skills, agents, hooks, MCP servers, and tools all ship in a consistent, versioned layout that's easy to automate around.
An Agentic Resource Discovery catalog lets other agents and registries find these capabilities programmatically.
From abstract math to .NET generics to APM authoring — each skill packages battle-tested, scoped domain knowledge.
MIT licensed, fully public source. Read exactly what your agents load — nothing hidden, nothing compiled away.
CI regenerates the marketplace manifest and discovery catalog on every change, so consumers always see the latest state.
Every skill's automation scripts are normalized to Python — no mixed toolchains to install, so behavior stays deterministic across every consumer.
agent-authoringFifteen skills covering every stage of building AI capabilities: author a single skill or agent, choose the right model for it, orchestrate several into a squad, embed and automate them, search the codebase, build portable agent knowledge and polished terminal interfaces, troubleshoot the harnesses that run them, then polish whatever Markdown they generate. Follow their built-in best practices and you get skills that are both more token-efficient to load and reference, and higher quality — well-triggered, unambiguous, and correctly structured.
Codifies compact writing rules, frontmatter constraints, and trigger-phrase description patterns so new skills are discoverable and unambiguous from the first draft.
The open, tool-agnostic format for giving coding agents project context — root and nested files, discovery rules, and migration from other formats.
Frontmatter, single-role prompts, and auto-delegation via well-written descriptions — the same trigger-quality bar as skills, one skill per harness:
Creates .agent.md agents for Copilot CLI — frontmatter fields, tool restrictions, and description-driven auto-delegation.
Creates .claude/agents/*.md subagents for Claude Code — frontmatter fields, tool restrictions, and delegation-phrasing conventions.
Chooses and tunes which AI model a custom agent, subagent, or squad node should use — capability tier, cost, speed, prompt-cache cadence, effort tuning, and cross-provider diversity — then maps the choice onto each harness's own field syntax.
Design multi-agent squads as a directed graph — sequential, parallel, branching, feedback-loop, and adversarial tribunal-review topologies — then delegate every node's actual file to whichever harness skill owns it.
Embed the Copilot agent runtime in your own app: sessions, custom tools, hooks, sub-agent delegation, and skill/plugin directories.
Run Copilot CLI safely in workflows — minimal tool allowlisting, auth strategy, and prompt-injection-aware permission scoping.
Searches files and code with fast literal, regex, glob, and file-type filters, while preferring a fresh code index for structural questions such as symbols, references, call paths, and impact analysis.
Finds, queries, redacts, and correlates logs, transcripts, and OpenTelemetry traces across Copilot in VS Code, Copilot CLI, Claude Code, Codex, Gemini CLI, and OpenCode — zero-token, script-driven queries instead of pasting raw logs into a prompt.
Produces, consumes, validates, and navigates Open Knowledge Format bundles: portable Markdown concept trees with progressive disclosure, linked relationships, provenance, trust, and lifecycle metadata.
Builds polished Python terminal interfaces and skill scripts with Textual and Rich, pairing responsive widgets and accessible keyboard workflows with deterministic --no-tui output for agents and automation.
Heading hierarchy, list and code-fence conventions, and table formatting so every generated document renders correctly everywhere — the other skill-and-agent-authoring skills all run it before calling a file finished.
Every stable Mermaid diagram type for conveying structure in far fewer tokens than prose — architecture, flows, and schemas included.
Converts JSON/YAML payloads to Token-Oriented Object Notation, cutting the tokens spent on tabular data inside prompts and skill references.
Install the whole plugin, or pull individual skills as you need them.
Ten plugins, ready to install. Each links to its source in the repository.
math-foundationsv0.1.0Released Aug 4, 2026
Core abstract math skills for angles, trigonometry, vectors, and coordinate system conversions.
math-rotationv0.1.0Released Aug 4, 2026
Abstract math skills for 3D rotation representations, quaternion algebra, and interpolation on manifolds.
math-geodesyv0.1.0Released Aug 4, 2026
Abstract math skills for geodetic coordinates, reference ellipsoids, and datum transformations.
dotnetv0.8.0Released Sep 11, 2026
C# and .NET skills for modern language best practices, generic math over System.Numerics, SIMD vectorization, enumerable/LINQ iteration performance, System.Threading.Channels producer-consumer pipelines, BenchmarkDotNet benchmarking, solution/project setup (Directory.Build.props, central package management, SourceLink), file-based apps (dotnet run file.cs, colon directives), and deterministic guardrails for AI code agents (analyzers, hooks, Aspire, testing).
dotnet-rxv0.1.0Released Sep 11, 2026
Rx.NET and ReactiveUI skills for composable observable pipelines, concurrency, lifetimes, MVVM state, commands, bindings, and deterministic testing.
apm-authoringv0.1.0Released Aug 4, 2026
Skills for authoring, consuming, and publishing APM packages, plus configuring apm-action CI workflows.
agent-authoringv0.15.0Released Aug 23, 2026
Skills for authoring AGENTS.md project-context files, SKILL.md agent skills, GitHub Copilot CLI custom agents, Claude Code custom subagents, compact and executable agent-facing Markdown, graph-pattern multi-agent squads/swarms, choosing which AI model an agent should use, building apps with the GitHub Copilot SDK, automating Copilot CLI in GitHub Actions, converting data to/from TOON (Token-Oriented Object Notation) to cut LLM token usage, authoring Mermaid diagrams in markdown files, following general Markdown formatting best practices, searching files and code with ripgrep, producing and consuming Open Knowledge Format bundles, creating polished Python terminal interfaces and skill scripts with Textual and Rich, and finding, querying, and troubleshooting logs, transcripts, and OpenTelemetry traces from coding-agent harnesses.
agentic-discoveryv0.1.0Released Aug 4, 2026
Skill for the Agentic Resource Discovery (ARD) open specification for publishing and discovering AI capabilities.
ucpv0.1.0Released Aug 5, 2026
Skills for implementing (business/merchant) and consuming (platform/agent) the Universal Commerce Protocol (UCP) for agentic commerce discovery, checkout, and payments.
codebase-memoryv0.1.2Released Aug 23, 2026
Skill for installing, verifying, and integrating codebase-memory-mcp (CBM), a local knowledge-graph code intelligence server, into agents, skills, and prompts that operate on a codebase, replacing grep/search/file-read exploration with structural graph queries via its CLI.
No plugins match your filters.
Intelligentium publishes an ARD v0.9 catalog so agents and registries can discover its plugins and skills programmatically.
https://intelligentium.ai/.well-known/ai-catalog.json