By alexei-led
Research, docs lookup, utility routing, skill authoring, instruction review, reasoning, and agent configuration audits
Sole mutator role — applies and verifies code, test, doc, and infra changes. Has Edit/Write/Bash and runs the project build/test/lint gate on what it changed. Use for implement, fix, refactor, or apply tasks. Not for read-only review (reviewer) or risk advice (advisor).
Read-only adversarial evaluator — reviews, audits, locates, or plans. Inspects with Read/Grep/Glob and does not modify code, run builds, or execute commands. Use for code review, security audit, locating code, or planning. Not for applying changes (engineer) or strategic risk verdicts (advisor).
Fast utility lane for simple bounded tasks on a cheaper model when available. Use proactively for file lookup, grep/glob searches, `git status/log/show/diff`, file reads, log summaries, and focused shell inspection. Not for code changes (engineer), adversarial review (reviewer), or strategic judgment (advisor).
Brainstorm ideas and stress-test draft plans before coding. Use when brainstorming, exploring approaches, designing a feature/API/flow, grilling or debating a bounded plan, challenging assumptions, or resolving design-blocking terminology. NOT for implementation task breakdown. NOT for generic technology comparisons or best-practice research; use researching-web. NOT for docs updates; use documenting-code.
Audit and improve AI coding-agent configuration. Use when reviewing or changing Claude Code, Pi, Codex, skill, agent, hook, MCP, permission, package, or generated-export setup. Default is review-only; fixes require explicit user approval or --fix. NOT for score-only instruction review or prompt lint; use reviewing-instructions. NOT for application config, git hygiene, code bugs, ordinary docs, or generated files without their source.
Find exact, version-correct library/API/framework docs through one lookup workflow. Use when the user says "look up docs", "how to use", "API for", "syntax for", "examples of", "show me the docs", mentions "ctx7"/"Context7", passes a `/org/project` library ID, or needs API signatures, config keys, syntax, examples, or versioned docs. NOT for comparisons, current-state or release-behavior questions, best-practice surveys, or recent ecosystem news — use researching-web.
Web research via platform web tools. Use for technical comparisons, current-state and release-behavior questions, recent facts, ecosystem news, best practices, standards, or questions needing grounded web evidence. NOT for exact API syntax, config keys, or code examples — use looking-up-docs for those. NOT for repo-specific questions — search local files first.
Use when asked to lint, audit, review, or score AI-facing instruction files such as SKILL.md, AGENT.md, AGENTS.md, CLAUDE.md, platform body.md files, prompt files, rules, policies, and agent-facing references. NOT for plugin manifests, application code review, harness configuration review, ordinary docs, tests, or generated build output.
Uses power tools
Uses Bash, Write, or Edit tools
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.
Portable skills, agents, hooks, and Pi-native extensions for Claude Code, Codex CLI, GitHub Copilot, Cursor, Grok, and Pi. Gemini is retired.
Requires Agent Bundler with build, check, and package, plus support for
per-agent sidecars, declared hook environments, Pi-native assets, and Codex
project-agent profiles. Verify packaging support:
agbun package --help
Build and check the complete generated tree:
make build
make check
make build is a direct agbun build --root . invocation. make check runs
the non-mutating agbun check --root . drift gate. Agent Bundler replaces the
configured dist/ output tree during builds, so do not edit generated files.
agentbundle.json # Agent Bundler manifest
src/
├── skills/<name>/SKILL.md # canonical YAML-frontmatter skills
│ └── .agentbundler/targets/*.json # optional target overlays
├── agents/<name>.md # canonical agent assets
├── .agentbundler/packages/*.json # package membership and metadata
├── hooks/<name>/ # typed Agent Bundler hooks
└── plugins/pi/<asset>/ # declarative Pi-native extension tree
└── extensions/ # Pi-native tools and compatibility runner
dist/<target>/<package>/ # generated Agent Bundler package roots
.claude-plugin/marketplace.json # repository-root compatibility wrapper
.agents/plugins/marketplace.json # repository-root compatibility wrapper
.codex/agents/*.toml # fixed-path Codex project-agent profiles
.github/plugin/marketplace.json # repository-root compatibility wrapper
.cursor-plugin/marketplace.json # repository-root compatibility wrapper
package.json#pi # merged Pi root compatibility manifest
Skills use Agent Bundler overlays for target-specific frontmatter, body, and
support-file changes. Target-wide preambles are composition entries in
agentbundle.json. Canonical source remains readable Markdown with YAML
frontmatter; Agent Bundler emits normalized JSON frontmatter.
Package membership is defined by src/.agentbundler/packages/*.json. Each
package lists exact skill and agent asset paths. Root marketplace files are
routing wrappers only: their entries point into dist/<target> and contain no
independent package trees. Do not edit dist/ or give the wrappers independent
metadata.
| Target | Output | Package contract |
|---|---|---|
| Claude Code | dist/claude/<package>/ | Marketplace packages, agents, and hooks |
| Codex CLI | dist/codex/<package>/ | Plugin packages plus .codex/agents/*.toml profiles |
| Pi | dist/pi/ | Aggregate package, typed hooks, and native extensions |
| GitHub Copilot | dist/copilot/<package>/ | Marketplace packages, agents, and hooks |
| Cursor | dist/cursor/<package>/ | Marketplace packages, agents, and supported hooks |
| Grok Build | dist/grok/<package>/ | Marketplace packages, agents, and hooks |
| Gemini CLI | none | Retired |
agbun package --root . --out DIR creates deterministic, target-native release
archives:
cc-thingz-claude.tar.gzcc-thingz-codex.tar.gzcc-thingz-pi.tgzcc-thingz-copilot.tar.gzcc-thingz-cursor.tar.gzcc-thingz-grok.tar.gzThe release workflow attaches all six files. Each archive expands directly to
the native package or marketplace root shown above. The Codex archive also
includes separate .codex/agents/*.toml profiles; they are not plugin contents.
Installation and marketplace registration use the target vendor's CLI; Agent
Bundler does not mutate user configuration or publish to a vendor registry.
Compatibility manifests at the repository root preserve Git and local-path
installation while keeping dist/<target> canonical:
Git worktrees, branch cleanup, hooks, secret scanning, config hygiene, and guardrails
Lightweight spec loop: plan one slice, execute one task, checkpoint or close
Code review, fixes, commits, and linting for development workflows
Browser-based testing, validation, screenshots, recordings, and quick automation
Idiomatic development across C# /.NET, Go, Java/Kotlin, Python, Rust, TypeScript, shell, and web
npx claudepluginhub alexei-led/cc-thingz --plugin discoveryEditorial "Skill Author" bundle for Claude Code from Agentic Awesome Skills.
Search, discover, and install AI agent skills from 9 registries with quality scoring, security labels, and cross-platform support
Skill and agent authoring tools: generate, audit, and improve Claude Code skills, commands, and agents
Evidence-based agent skills compiler with progressive capability tiers (Quick/Forge/Forge+/Deep).
Plugin-safe Claude Code distribution of Antigravity Awesome Skills with 1,419 supported skills.
Toolkit for authoring and validating Claude Code skills and agents. Script-backed skills-toolkit skill runs a real deterministic validator (validate_skills.py), plus agent-author/agent-validator and 13 templates. Auto-invoke on 'create skill', 'validate agent', 'build agent'.