Agent Plugins Marketplace

Recently added

Plugins grouped by the day they were added to the directory, newest first.

Sep 22, 2026

  1. mnemon0stars

    Research oracle: an agent that holds large documentation or an external repo in its own context and answers your follow-up questions with citations, optionally scoped by a graphify graph

    CodexClaude Code1 skill
  2. wigtn-plugins-with-codex1stars

    Selective product contracts and visual craft for Codex: PRDs, saved WorkGraphs, screen artifacts, UI and document composition, acceptance evidence, and safe release actions.

    Codex10 skills
  3. wigtn-knowledge-wiki1stars

    Opt-in, gated Codex session knowledge capture for an explicitly authorized personal wiki namespace.

    Codex1 skill
  4. wigtn-plugins-with-codex1stars

    Selective product contracts for Codex: compact PRDs, saved WorkGraphs, scoped screen artifacts, acceptance evidence, explicit verified delivery, and safe release actions.

    Codex9 skills
  5. wigtn-plugins-with-codex1stars

    Selective product contracts for Codex: compact PRDs, saved WorkGraphs, scoped screen artifacts, acceptance evidence, explicit verified delivery, and safe release actions.

    Codex9 skills
  6. wigtn-plugins-with-codex1stars

    Selective product contracts for Codex: compact PRDs, saved WorkGraphs, scoped screen artifacts, acceptance evidence, explicit verified delivery, and safe release actions.

    Codex9 skills
  7. wigtn-plugins-with-codex1stars

    Selective product contracts for Codex: compact PRDs, saved WorkGraphs, scoped screen artifacts, acceptance evidence, explicit verified delivery, and safe release actions.

    Codex9 skills
  8. jev-decision0stars

    Build AI-powered software with TypeSafe System One and Jev using typed judgments, calibrated probabilities, and composable decision primitives.

    Codex
  9. jev-decision0stars

    Build AI-powered software with TypeSafe System One and Jev using typed judgments, calibrated probabilities, and composable decision primitives.

    Codex
  10. sceneview1.3kstars

    Build 3D and AR apps with the SceneView SDK — Jetpack Compose + Filament + ARCore, SwiftUI + RealityKit, and Filament.js on the web. Three skills plus the free sceneview-mcp server.

    Codex1 MCP server
  11. simple-vlog-editor0stars

    Understand and edit complete videos in the visible SimpleVlogEditor through its local MCP server.

    Codex2 skills1 MCP server
  12. eridian20stars

    Rocky-speak for Codex — terse Eridian token compression, workflows, stats, and buddy controls

    CodexClaude Code1 skill
  13. apartment-ranking0stars

    Apartment ranking and architectural document ingestion workflows.

    CodexAgent Plugins2 skills
  14. sot-runtime0stars

    External State-of-Thought controller for Astra/Codex.

    CodexClaude CodeAgent Plugins1 skill1 MCP server
  15. driftnote0stars

    Capture browser sources into a local Obsidian vault, configure safe formatting for the resulting web clips, and turn a captured note into a brainstorm-style Excalidraw diagram.

    CodexClaude CodeAgent Plugins3 skills
  16. github-radar0stars

    Read-only daily rankings for new and fast-growing public GitHub repositories.

    Codex1 skill
  17. agent-bus0stars

    Puts this session on the agent bus: a shared message bus between coding agents and any other participant that follows PROTOCOL.md. Ships the MCP server connection (send, inbox, ack, who, plus the agent-bus://protocol and agent-bus://inbox/{address} resources), a skill on how to use it, and two hooks. SessionStart prints 'you are claude@<repo>#<session>' so an agent that has lost context can tell its own rows from a peer's; both hooks then run the inbox bridge, which prints unread messages addressed to that instance as plain text with the ack command to run after reading (it does not ack itself, since a prompt queued mid-turn drops hook output); it is silent when there is no mail or the daemon is unreachable, so it costs nothing per turn unless something arrived. The daemon itself runs once on a host machine; see README.md.

    Claude Code1 skill
  18. arkspace0stars

    Provider-aware Web retrieval, structured extraction, cited Research, Code Context, and reusable agent skills.

    CodexClaude Code5 skills
  19. application-agent-kit0stars

    A privacy-first system for setting up, preparing, reviewing, tracking, and scheduling monitoring of personal applications.

    Codex5 skills
  20. bymax-workflow0stars

    Phased planning + execution workflow for Claude Code. Spec → Roadmap → Phase-Tasks → Task, each step with explicit user-approval gates and JIRA-style dashboards. Includes /bymax-workflow:brainstorm, /bymax-workflow:plan, /bymax-workflow:verify, /bymax-workflow:checkpoint, the universal /bymax-workflow:standards reference (with the §0 reuse-first simplicity ladder), and /bymax-workflow:autopilot — the loop-engineering executor that drives an approved roadmap to done autonomously, one merge-gated PR per phase.

    Claude Code2 skills
  21. bymax-web-verify0stars

    Real-browser verification and assisted UI testing for web apps. /bymax-web-verify:test brings the full stack up — reuses a running backend instance or starts one, opens the frontend in the Claude Desktop Browser pane (preview) — and walks a UI flow step by step while the user watches, verifying via page state, console, network (2xx gate), and server logs; falls back to the agent-browser CLI in a terminal. /bymax-web-verify:verify drives a real browser to confirm a single change works. /bymax-web-verify:setup installs the agent-browser CLI + Chrome for Testing in one shot. A SessionStart hook detects a missing CLI and offers to install it. Depends on `agent-browser` but never bundles it. /bymax-web-verify:record captures a UI flow as reviewable video evidence: the project's own Playwright with video forced on for one isolated run, human pacing (slowMo with a resized timeout), ffmpeg lead-in trim with a verified first frame, an ffprobe pacing gate, optional --mp4, and a plain-text walkthrough derived from the spec that ran.

    Claude Code
  22. bymax-report0stars

    The weekly silent-standup report for one repository, written from evidence. /bymax-report:standup collects what shipped in a period (pull requests via gh, commits via git) and what was asked (what a person typed into Claude Code or Codex sessions on that repository, worktree sessions included, pasted third-party requests included), then writes the two sections the team reads: a PROGRESS list of the asks as outcome sentences, undelivered ones marked in progress, and UPDATES grouped by product area saying what changed and what it replaced. No hashes, PR numbers or file names reach the report; an evidence block under it maps every line back to its PR, commit and request for the author. Plain text printed to paste, no Markdown, no file left on disk, English by default; --author keeps one git author. Depends on git; gh is used when present and its absence is reported, never hidden.

    Claude Code1 skill
  23. bymax-quality0stars

    Evidence-based quality gates: bounded Claude and Codex review with pinned candidates, verified findings, correction deltas and regression checks; TDD, stack-specific reviewers and security hooks.

    Claude Code1 skill
  24. bymax-qa0stars

    Whole-system QA and security audit for any project, with evidence and hand-off. Point /bymax-qa:audit at a target: a Jira ticket key (it verifies Jira access, reads the acceptance criteria, and reports each PASS/FAIL/BLOCKED/NOT-VERIFIABLE with evidence), a branch or ref range or PR (it scopes the hunt to the change, like a code review), or nothing (the whole system via a signed scope). It maps the stack and its trust boundaries, hunts by domain (authentication, authorization and tenant isolation, injection and input, cache/Redis, database/Postgres, observability, transport and config, architecture, frontend, supply chain) with read-only finder agents, probes the running stack against the allow-listed hosts only, and admits a finding only after verification: a runnable reproduction, captured evidence and verified impact, with an independent verifier for every hunted candidate. Every finding carries ASVS 5.0, CWE and API Top 10 references, a CVSS vector, and a status that moves OPEN → HANDED-OFF → FIX-CLAIMED → VALIDATED or REOPENED. Fixes are never applied by the auditor: findings are handed to the owning peer session over cross-session messaging, filed as GitHub issues when no peer is live, or written back to the ticket as a comment, and re-tested before they close. A PreToolUse hook is a floor under the scope while an audit is active: a network tool in command position may only reach allow-listed hosts, and a Write/Edit tool call may only land in .claude/qa/ (arbitrary Bash writes and unlisted network binaries rest on the skill's never-edit-the-target discipline and its read-only finders). External scanners (semgrep, gitleaks, osv-scanner, trivy, zap, nuclei, axe, Lighthouse) are used when present and never bundled.

    Claude Code
  25. bymax-pr0stars

    The PR lifecycle, powered by the `gh` CLI. /bymax-pr:push ships work safely: branch (never the default branch) → stage → complete Conventional-Commits message → push, with an explicit `pr` opt-in that opens a fully-described PR (Summary / Changes / How to verify). /bymax-pr:babysit-pr autonomously shepherds an open PR to merge-readiness: resolves merge conflicts, polls CI at a cadence matched to the CI's real duration (270 s floor), classifies failures as real vs flaky (re-running flaky checks), fixes real failures locally before pushing, triages bot review comments (4-tier), and notifies you when the PR is green. Never merges, never pushes to the base branch, never force-pushes. A built-in preflight verifies the gh CLI is installed and authenticated. Depends on `gh` + `git` but never bundles them.

    Claude Code1 skill
  26. bymax-pm0stars

    Engineering Project Manager for multi-agent development. The /bymax-pm:pm command turns the current session into a PM/TPM that coordinates INDEPENDENT Claude Code peer sessions: discovers agents via ListAgents, delegates structured task contracts via SendMessage, tracks a deterministic task lifecycle (assignment → acknowledgement → evidence → independent review → PM verification), detects stuck agents through one-shot idle notices instead of polling, resolves blockers between agents before escalating, and keeps all project truth in a persistent, git-friendly .claude/pm/ workspace that survives session restarts. Evidence-driven: a worker saying done is never enough — DONE requires diffs, tests, CI and PM verification. The PM coordinates; it does not implement.

    Claude Code1 skill
  27. bymax-mobile0stars

    Slash commands to boot the iOS Simulator (/bymax-mobile:sim-ios) and Android Emulator (/bymax-mobile:sim-android) for Expo / React Native projects. Auto-detects whether `expo start` (Metro reattach) or `expo run:ios|run:android` (full rebuild + install) is the right call. macOS-first; Android also works on Linux when the SDK is configured.

    Claude Code
  28. bymax-bootstrap0stars

    Scaffold new projects with strict TypeScript, ESLint flat-config (security + import-order + suppression bans), Prettier, .vscode (format-on-save), Husky + commitlint + lint-staged, and CLAUDE.md / AGENTS.md / README.md templates. Detects the stack (Next.js, Expo/RN, Vite+React, Node) and uses the right ESLint preset. Includes /bymax-bootstrap:upgrade-standards for non-destructive upgrades on existing projects.

    Claude Code
  29. bymax-all0stars

    Meta-plugin reference for the complete bymax toolkit (workflow + quality + bootstrap + mobile + web-verify + pr + pm + qa + report). Note: Claude Code's plugin manifest does not auto-install dependencies — install bymax-workflow, bymax-quality, bymax-bootstrap, bymax-mobile, bymax-web-verify, bymax-pr, bymax-pm, bymax-qa, and bymax-report individually.

    Claude Code
  30. bymax-codex0stars

    Bymax workflows for Codex: evidence-based code review, planning, testing, PR management and capability-aware orchestration.

    Codex28 skills
  31. embedded0stars

    显式选择后启用的公开版嵌入式开发工具箱，涵盖 STM32、MSPM0、ESP32、数字电源、FOC 与电赛等模块。

    Codex1 skill
  32. test-sloc-cut22stars

    Shrink a test suite without losing an asserted fact or a covered branch: a fact matrix of what each test pins plus per-test branch coverage and a verified covering set, deleting only on the intersection, with a mutant check for every relocated fact. Use when asked to trim, dedupe, consolidate, or reduce tests or test SLOC, when a suite grew by accretion, when deciding whether to delete, fold, or parametrize tests, or before cutting a test on 'coverage is unchanged' alone. Python/pytest.

    Codex1 skill
  33. python-complexity22stars

    Measure Python complexity per function and per call path with ruff and complexipy; optionally add Jev judgments through configurable providers (Vercel by default). Use when prioritizing refactors, investigating over-layering, or verifying that a refactor improved the whole path. Without the selected provider key, semantic review is skipped and static analysis continues. Tests verify behavior.

    Codex1 skill
  34. jev22stars

    Consult Jev for bounded typed second opinions during agent work: compare plans, check claims against supplied evidence, prioritize investigations, and challenge conclusions. Uses Vercel AI Gateway by default with configurable providers; skips when the selected key is absent. Includes a standalone uv-run helper and offline contract tests.

    Codex1 skill
  35. doc-quality22stars

    Review and improve engineering Markdown designs, analyses, implementation plans, and ADRs. Measure sections and protected content with a Markdown parser; optionally use Jev for typed quality and meaning-preservation judgments. Preserve requirements, uncertainty, code, and links while the agent makes targeted edits. Jev is optional and provider configurable.

    Codex1 skill
  36. gborges-standard0stars

    Standard Codex setup used across all of Gabe Borges' repos. Injects the Plain English rules at session start and into every subagent, strips AI-attribution footers from GitHub writes, re-injects the writing rules reminder with every user message, and ships the writing-voice, read-aloud-prep, bear-notes, codex-delegate, model-routing-review, and pair-debate skills, plus grill, build, spec, routine, wayfinder, handoff, research, wait-what, teach, writing-for-agents, domain-modeling, investigate, and resolving-merge-conflicts.

    CodexClaude Code18 skills
  37. invisible-playwright-mcp31.6kstars

    AI browser agent: browses, clicks, types, and reads real web pages from plain-English instructions.

    CodexClaude CodeAgent Plugins1 skill1 MCP server
  38. hatch0stars

    Drive Hatch, the browser on this Mac whose live pages the user watches, from any project.

    CodexClaude Code1 MCP server
  39. planwerk0stars

    Guided, interactive skills for the planwerk-agent workflow

    Claude Code10 skills
  40. ai-power-apps0stars

    Upgrade the AI you already use. One connection gives it live data across travel, markets, research, reading, code, food, fitness and social — with real follow-ups and real links — plus a performance stage, Stock to Stage 3+, that puts more behind every answer. You choose what's on.

    Claude Code1 skill1 MCP server
  41. cognitive-biases0stars

    Доказательный советник по когнитивным искажениям: 111 искажений в пяти тирах по размеру эффекта и репликации, структурный Bias Scan ситуации, протокол дебайсинга по каждому искажению, разбор популярных мифов.

    Claude Code1 skill
  42. soulmate0stars

    It verifies what you asked an agent to do and what came back, in the same record. Use it as an opt-in, bounded project guidance surface; ordinary work remains unchanged. Plugin installation provides host guidance and does not install the separate soulmate CLI.

    CodexClaude Code
  43. exitbind0stars

    Exitbind — No result exits unbound. Bind the evidence to the exact result before exit. For your existing Codex or Claude lead; this skills-only plugin installs guidance, not the separate Exitbind CLI.

    CodexClaude CodeAgent Plugins
  44. exitbind0stars

    Exitbind — No result exits unbound. Bind the evidence to the exact result before exit. For your existing Codex or Claude lead; this skills-only plugin installs guidance, not the separate Exitbind CLI.

    Agent Plugins2 skills
  45. agent-dispatcher46stars

    Routes each request to one of 27 specialist agent roles, which load only the skills the work needs and verify what they claim.

    Claude Code1 skill
  46. midnight-vibe0stars

    A harness that cannot start without interviewing you and cannot finish without evidence. The phase lives in one .midnight/prd.md and hooks guard its transitions.

    Claude Code1 skill
  47. marathon0stars

    Sustainable long-haul development workflow with context engineering: concept-driven init, code and context project kinds, plan/start/experiment sessions with review and docs passes, staged execution with per-stage review, standalone-repository or workspace operation with cross-repo steps, mid-session handoff/resume, a standardized session pipeline with universal extension hooks, and volatile-vs-stable context promotion and decay

    Claude Code
  48. marathon-roadmap0stars

    Marathon extension keeping a roadmap manifest — goals, tasks, and a backlog in context/roadmap.toml — current through marathon's session hooks: conventions layered in at session start, dotted task citations in the session record, the finished task deleted and next advanced at closeout

    Claude Code1 skill
  49. marathon-architecture0stars

    Marathon extension adding the architecture layer — the principles, definitions, and conventions that have generalized past one repository — as a standalone project's top-level architecture/ directory or a workspace's designated architecture repository, filled by promotion from design/ through marathon's session hooks

    Claude Code1 skill
  50. splash0stars

    Splash: Visual journalism for every newsroom — a skill pack a newsroom installs and drives from its own AI assistant. Entry point: the splash skill.

    Claude Code16 skills