{"data":[{"slug":"preserve-session","title":"preserve-session","description":"Preserves Claude Code session history across project directory renames, moves, and copies by assigning each project a path-independent UUID and maintaining a global registry mapping UUID to current path.\n\nAddresses a long-standing pain point: --resume and --continue silently lose session history when paths change. Closes anthropics/claude-code#5768; also tracked in #38089, #28745, #26766, #27883.\n\nCommands: /preserve-session:fix recovers sessions after rename/move (with copy detection and automatic .jsonl merge on destination collision), /preserve-session:inherit copies session history from another registered project, /preserve-session:doctor diagnoses hash/registry state and slug collisions including non-ASCII path issues on macOS, /preserve-session:uninstall removes all preserve-session data with a preview-then-confirm safety pattern. A SessionStart hook auto-initializes .claude/hash.txt on first run — zero manual setup.\n\n  Demo: https://www.youtube.com/watch?v=lWMrzASG7_o\n\n  Quality signals: 24 verified test scenarios (rename, move, copy, slug collision, non-ASCII paths, registry corruption recovery, full --resume E2E). Atomic registry writes via fcntl.LOCK_EX + tempfile.mkstemp + os.replace. NFC slug normalization to match Claude Code's path encoding on macOS. Python heredoc isolation against shell injection. Symlink-safe uninstall (isinstance/dict guards).\n  Shared helpers in hooks/common.sh with cross-platform Python discovery and uuidgen fallback.\n\n  Full design rationale, security notes, and test plan are in two related submissions:\n  • https://github.com/anthropics/claude-code/pull/39148 (initial bundled-plugin attempt; comprehensive design + 24-item test plan with collapsible sections)\n  • https://github.com/anthropics/claude-plugins-official/pull/1347 (auto-closed by submission bot; retained as reference for the proposed marketplace.json entry)\n\n  Plugin location: lives at plugins/preserve-session/ inside the wonbywondev/claude-plugins monorepo. The marketplace entry should use the git-subdir source format pointing to that subdirectory.","developer":null,"category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/wonbywondev/claude-plugins.git"},{"slug":"praxis-engineering","title":"praxis-engineering","description":"8 flagship engineering skills curated from the Claude Skills library, covering the full shipping flow: architecture (senior-fullstack), delivery (senior-devops, docker-development, terraform-patterns), security (senior-security), AI tooling (mcp-server-builder, rag-architect) and code review. Each skill is a self-contained folder with SKILL.md instructions, Python automation scripts (standard library only, no pip installs), reference knowledge and examples. Install once, get production engineering expertise across 8 domains.","developer":null,"category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"borghei/Claude-Skills"},{"slug":"pearl-llm-memory","title":"pearl-llm-memory","description":"Externalized Cognitive State Transfer\n\nPEARL-LLM Memory (.pearl) is a JSON-based file format for storing structured, layered, verifiable context that persists across agent sessions. It gives autonomous agents a durable memory with immutable history, temperature-graded recall, cryptographic chain of custody, and branching decision trails.\n\nThe format is agent-agnostic. The reference implementation is a zero-dependency Python CLI. Both are open.","developer":null,"category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/reneemgagnon/Pearl_LLM_Memory.git"},{"slug":"orchestra","title":"orchestra","description":"Disciplined AI engineering toolkit for Claude Code. Ships typed design-doc templates (Feature LLD, Bug Report, ADR, Postmortem, Runbook, Design Doc) with a 4-gate spec review (Completeness / Evidence / Clarity / Consistency), mandatory mermaid diagrams, an octal-safe auto-numbering script, AGENTS.md and llms.txt sync, a bug-iteration loop that prevents orphan fix-commits, and an enforced doc-before-code workflow with five named gates (Discovery / Design / Spec Review / Commit / Implementation Sync). Includes a Python lint CLI for Refs:-line + metadata + status validation, an auto DECISIONS.md generator with bidirectional ADR supersession consistency. Patterns adopted from Michael Nygard, Google SRE, Pragmatic Engineer, agents.md, and llmstxt.org. First skill of the Orchestra umbrella; future skills cover workflow, registry, tasks, gates, and plans.","developer":"Hassan Mohiddin","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/hassan-mohiddin/orchestra.git"},{"slug":"open-tax","title":"open-tax","description":"Free, open-source US federal tax preparation assistant for Claude Code. Drop your W-2s, 1099s, and crypto exchange CSVs into the workspace — Open Tax parses them, computes your federal tax liability with full bracket breakdowns, handles capital gains (Form 8949), backdoor Roth IRA (Form 8606), excess Social Security tax, and crypto staking income. It runs sanity checks, suggests missed deductions, and exports ready-to-import files for FreeTaxUSA or TurboTax. 100% local and private — no tax data ever leaves your machine. Zero pip dependencies — Python stdlib only.","developer":"Tugberk Goc","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/tugberkgoc/open-tax.git"},{"slug":"nlpm","title":"nlpm","description":"NLPM lints, scores, fixes, and trend-tracks the natural-language artifacts that make up a Claude Code plugin — skills, agents, commands, rules, hooks, manifests — against a 50-rule rubric on a 100-point scale. Ships focused agents (scanner, scorer, checker, fixer, tester, security-scanner, vague-scanner, vocab-drift-scanner) and a standalone Python validator (bin/nlpm-check) that runs in pre-commit hooks and CI without Claude Code installed, plus a shields.io \"Validated by NLPM\" badge.","developer":"xiaolai","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/xiaolai/nlpm-for-claude.git"},{"slug":"neo-skills","title":"neo-skills","description":"提供 .NET、Azure Pipelines、程式碼審查、Git 工作流程與 Swift/Python 開發支援","developer":"Benknightdark","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/Benknightdark/neo-skills.git"},{"slug":"mine","title":"mine","description":"Mines every Claude Code session into a local SQLite database and provides a /mine skill with 18+ natural language query intents. Track costs, tokens, tools, projects, errors, cache efficiency, and full-text search across all your past conversations. Features automatic session ingestion via hooks, error memory that warns when you repeat mistakes, cost anomaly detection, project/global scope detection, and API value tracking with subscription ROI comparisons. Zero external dependencies (stdlib Python only)","developer":null,"category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"anipotts/claude-code-tips"},{"slug":"memoriant-test-coverage-skill","title":"memoriant-test-coverage-skill","description":"AST-powered test coverage analysis for Claude Code. Scans source and test directories, identifies untested functions by parsing Python abstract syntax trees, and generates pytest test skeletons automatically. Finds coverage gaps that line-based tools miss by understanding function signatures, class methods, and module structure.","developer":"Nathan Maine","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/NathanMaine/memoriant-test-coverage-skill.git"},{"slug":"memoriant-eval-sandbox-skill","title":"memoriant-eval-sandbox-skill","description":"AI agent evaluation sandbox for Claude Code. Holdout scenario testing with Doer/Judge/Adversary/Observer roles, probabilistic satisfaction scoring, and append-only JSONL audit trails with integrity hashes. Test your agents before deploying them. Includes full Python evaluation framework.","developer":"Nathan Maine","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/NathanMaine/memoriant-eval-sandbox-skill.git"},{"slug":"mcpg","title":"mcpg","description":"Generate MCP server scaffold and tool implementation code from a natural language description.\n\nDescribe what tools you want in plain language, choose TypeScript or Python, and get a working server with implemented tool handlers — ready to install into Claude Code. Supports both direct mode (with arguments) and interactive mode (step-by-step prompts).","developer":null,"category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"yn01/claude-plugins"},{"slug":"lazyline","title":"lazyline","description":"Line-level Python profiler skill for Claude Code. Profile any package or script to find exactly which lines are slow, then optimize iteratively.","developer":"Tomáš Venkrbec","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/TomasVenkrbec/lazyline.git"},{"slug":"launchpad","title":"launchpad","description":"The most comprehensive agentic coding harness for Claude Code that works for both greenfield and brownfield projects. Drops 38 slash commands, 36 sub-agents, and 16 skills into any repo to give Claude Code full project context, run it through structured shape → plan → build → review → ship loops, and enforce multi-agent quality gates before any change reaches main. Works in any existing TypeScript, Python, Go, Django, or polyglot repo — install the plugin and run /lp-kickoff to start.","developer":"Foad Shafighi","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/builtform/launchpad.git"},{"slug":"karellen-jdb-mcp","title":"karellen-jdb-mcp","description":"JDB (Java Debugger) plugin for Claude Code. Attach to running JVMs via JDWP and debug interactively with breakpoints, watchpoints,\nexception breakpoints, stepping, expression evaluation, thread inspection, stack navigation, and class introspection.\n\n### Features\n\n- **Connect & Launch**: Attach to any JVM started with JDWP, or launch one directly with `jdb_launch`\n- **Multi-session**: Debug multiple JVMs simultaneously on different ports\n- **Breakpoints & Watchpoints**: Line/method breakpoints, field watchpoints, exception breakpoints with caught/uncaught filtering\n- **Execution Control**: Step in, step over, step out, continue, suspend/resume threads\n- **Inspection**: Evaluate expressions, dump objects, view locals, list loaded classes/methods/fields\n- **Stack Navigation**: Walk the call stack, move up/down frames, pop/re-enter frames\n- **Concurrency**: Thread listing, lock/monitor inspection for deadlock analysis\n- **Tracing & Monitoring**: Trace method entries/exits, auto-execute commands on each stop\n- **JDK Compatibility**: Supports JDK 8+ with feature detection (virtual threads on JDK 20+, thread-scoped breakpoints on JDK 13+)\n\n### Prerequisites\n\n- Python 3.10+\n- A JDK installation (for the `jdb` executable)\n\n### Installation\n\npip install karellen-jdb-mcp","developer":"Karellen, Inc.","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/karellen/karellen-jdb-mcp.git"},{"slug":"inkbox","title":"inkbox","description":"Inkbox gives Claude Code agents a persistent identity with a real email inbox, phone number, contacts, notes, and a zero-knowledge encrypted vault. The plugin bundles skills for the Inkbox TypeScript SDK, Python SDK, and CLI so Claude can use these capabilities directly from the editor.","developer":"Inkbox AI","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/inkbox-ai/inkbox.git"},{"slug":"hwpx","title":"hwpx","description":"Read, write, and convert HWP/HWPX (한글) documents — South Korea's dominant document format. Extract text, math equations, images, and tables from HWP/HWPX files. Generate HWPX documents programmatically with B4/A4 layouts, equations, and images. Convert between HWP/HWPX and Markdown bidirectionally. Includes HWP equation script syntax knowledge, HWPX XML schema reference, and ready-to-run Python converter scripts.","developer":"Hohyon Ryu","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/hohyon-ryu/claude-hwpx-plugin.git"},{"slug":"harness-engineering","title":"harness-engineering","description":"AI Agent Harness Engineering: bootstrap, audit, and evolve a reliable 6-layer agent harness for any project. Generates AGENTS.md, hooks (type-check, .env protection, auto-format), architecture docs, and sprint planning tools. Supports Claude Code and CodeBuddy. Works with TypeScript, Python, Go, Java, and generic stacks.","developer":"Harness Engineering","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/huangbaixun/harness-engineering.git"},{"slug":"guard","title":"guard","description":"PreToolUse hooks for Claude Code that deny dangerous shell, git, and credential commands before they run: `rm -rf /` shapes, `git push --force`, `python -c` / `node -e` eval, pipe-to-shell, live credentials (`gh auth token`, `aws sts get-session-token`, `op read`), `git -c core.hooksPath=…` injection, silent commit reuse (`git commit -C HEAD~1`, `--reuse-message`). Each deny names the rule and prints a one-line override (`guard allowlist allow-command …`); decisions stream to a JSONL log the optional CLI reads (`pipx install tracine-guard`, then `guard noisy --since 24h` or `guard trace <session>`). Apache-2.0, no third-party deps. Defense in depth alongside Claude Code's own permissions — especially valuable for autonomous agent windows where no human is present to answer an ASK, so denies act as a hard floor. Not a sandbox or exfiltration boundary; see SECURITY.md.","developer":"TracineHQ contributors","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/TracineHQ/guard.git"},{"slug":"gsc-pap","title":"gsc-pap","description":"Ask Claude about your Google Search Console top queries, decaying pages, ranking trends, content cannibalization, quick wins.\n\nPlug-and-play hosted MCP bridge: sign in with Google, copy a token, paste it. No GCP project, no Python, no JSON config files to manage.\n\n18 read-only tools covering Search Analytics, URL Inspection, Sitemaps, and a built-in archive for trending data beyond Google's 16-month window.\n\nPer-site authorization layer means you decide which properties Claude can see, even though OAuth grants account-wide access. All tool calls audited at app.gscpap.com\n\nBuilt for SEO consultants, in-house marketers, and developers who want to delegate analysis to AI without exposing their Google credentials or wrangling 1000-line CSV exports.","developer":"Ali Arif Soydas","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/aliarifsoydas/gsc-pap-claude-plugin.git"},{"slug":"graperoot","title":"graperoot","description":"Dual-graph code context engine for Claude Code. Builds a structural call graph and dependency graph of your codebase, then delivers precise file and symbol recommendations before Claude starts exploring. Reduces token usage by up to 60% and improves code accuracy by guiding Claude to the right files first. Supports TypeScript, JavaScript, Python, Go, Rust, Java, C/C++, and 10+ more languages. Install via pip, no source code required, fully compiled","developer":"Krishnakant","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/kunal12203/graperoot-plugin.git"},{"slug":"goodmem--community","title":"goodmem","description":"GoodMem memory infrastructure for AI agents. Use Python SDK skills to write code that manages embedders, spaces, and memories, or use MCP tools to perform GoodMem operations directly via natural language.","developer":"PAIR Systems","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/PAIR-Systems-Inc/goodmem-claude-code-plugin.git"},{"slug":"gomboc-ai-community","title":"gomboc-ai-community","description":"A Claude Code plugin for scanning, fixing, and creating ORL (Open Remediation Language) rules across Infrastructure as Code, containers, orchestration, and application code. Supports Terraform, HCL/Terragrunt, CloudFormation (YAML + JSON), Bicep, Dockerfile, Kubernetes, and Python.","developer":"Gomboc AI","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/Gomboc-AI/gomboc-community-skills.git"},{"slug":"gimme-the-lint","title":"gimme-the-lint","description":"Progressive linting system that only blocks on new violations, never on legacy code. Baselines existing lint issues by directory, then enforces clean code going forward. Supports Python (Ruff) and JavaScript/TypeScript (ESLint) in monorepo configurations. Includes drift detection, auto-healing, git hooks, and a GitHub Action for CI enforcement. Directory-chunked baselines allow gradual adoption without a big-bang cleanup.","developer":"TheGlitchKing","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/TheGlitchKing/gimme-the-lint.git"},{"slug":"fullstack-dev-skills","title":"fullstack-dev-skills","description":"66 specialized skills for full-stack development: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing skills. Includes 9 project workflow commands for epic planning, discovery, execution, and retrospectives.","developer":"Jeffallan","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/jeffallan/claude-skills.git"},{"slug":"formal-agent-contracts","title":"formal-agent-contracts","description":"Formal methods toolkit for multi-agent development. Define inter-agent contracts in VDM-SL, verify with VDMJ, prove with Z3, and generate TypeScript/Python code with runtime contract enforcement — all through natural language conversation with Claude.","developer":"IID Systems","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/kotaroyamame/formal-agent-contracts.git"},{"slug":"forge","title":"forge","description":"Description:\n\nAdversarial code hardening for Claude Code. Forge spawns two agents — a Builder and a Breaker — that iterate against each other until the code is unbreakable. The Builder implements a feature, the Breaker writes adversarial tests to prove bugs, and the Builder must fix the implementation without deleting or weakening the tests. The loop continues until the Breaker can't produce a failing test. The result is battle-tested code with a permanent adversarial test suite.\n\nKey features:\n\n  - Evidence-based, not opinion-based. Every issue the Breaker raises must have a failing test that proves the bug. No test, no break. This eliminates false positives and subjective feedback.\n  - Iterative adversarial loop. Builder and Breaker are separate agents that go back and forth for up to N rounds. The Builder cannot dismiss feedback; it must make the failing tests pass.\n  - Adversarial tests are permanent. The Breaker's tests stay in the repo under tests/adversarial/ as a regression suite. Future changes are protected by the same tests that originally caught the bugs.\n  - Automatic checkpoint. Creates a git tag before starting so the entire forge session can be rolled back with one command.\n  - Configurable rounds. Default 5 rounds, adjustable with --rounds N. Stops early if the Breaker can't find issues.\n  - Auto-detects project stack. Reads config files to determine language, test framework, and conventions. Works with Python, TypeScript, Rust, Go, Ruby, and anything with a test runner.","developer":null,"category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/UtkarshJain98/forge.git"},{"slug":"foresight-intelligence","title":"foresight-intelligence","description":"Strategic foresight intelligence using IFTF methodology. Two prediction modes: Soft Predict Future (Claude-native skill, instant, claude.ai + Claude Code) and Hard Predict Future (deterministic 12-step pipeline, Python-computed — Claude Code only). Both output four independent futures (Probable/Plausible/Possible/Preferable) with per-stakeholder conditional analysis. Try: Will India become the global AI leader? Will OpenAI or Anthropic win?","developer":"Santhosh Gandhi","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/isanthoshgandhi/foresight-engine.git"},{"slug":"foresight-engine","title":"foresight-engine","description":"Strategic intelligence plugin for Claude Code. Give it any real-world forecasting question and it produces a structured four-scenario analysis (Probable, Plausible, Possible, Preferable) backed by live web signals, historical analogues, and deterministic probability scoring. Built on IFTF foresight methodology: STEEEP framework, Futures Cone, Cross-Impact Analysis, and Backcasting. Two-layer architecture: Python handles all deterministic arithmetic; Claude handles web search, reasoning, and scenario writing.","developer":"Santhosh Gandhi","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/isanthoshgandhi/foresight-engine.git"},{"slug":"find-conversation","title":"find-conversation","description":"Search past Claude Code conversations by keyword and instantly find any session to resume. The tool scans your entire conversation history across all projects and outputs a formatted table with session slugs, project paths, timestamps, message previews, match context, and ready-to-copy resume commands.\n\nIt uses binary pre-filtering and single-pass parsing for fast results, with no external dependencies — compatible with Python 3.8 and above.","developer":null,"category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/akzarma/claude-find-conversation.git"},{"slug":"everyday-causal-skills","title":"everyday-causal-skills","description":"A causal inference plugin for AI agents. Describe a problem in plain language and it helps you choose a method, check assumptions, write the analysis in R or Python, and stress-test the results.\n\nWho this is for: Anyone who needs to measure whether something actually worked, such as Marketing and growth teams; Product managers and BI analysts; Data scientists; Revenue and operations teams; Policy researchers; Students and self-taught practitioners.\n\nWhat you get from this plugin\nThe plugin works in five steps, from refining the question you want to answer, to writing the report. You are free to pick and start from any step you like.\n\n→ Describe your problem\n→ Get a method recommendation\n→ Check assumptions and structure the analysis\n→ Stress-test the results\n→ Write the executive report","developer":"Robson Tigre","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/RobsonTigre/everyday-causal-skills.git"},{"slug":"eight-eyes","title":"eight-eyes","description":"Multi-agent code review with 8 hook-enforced roles. Each role targets a different failure surface — the skeptic reviews blind without the implementer's narrative, the security auditor can only run approved scan commands, the test writer cannot touch source code, and the implementer cannot run Bash. Scope boundaries are enforced through deterministic Python hooks at the PreToolUse layer, not through prompt instructions.\n\nThe 8 roles: implementer (scoped writes, no Bash), test-writer (test directories only), skeptic (blind read-only review), security auditor (read-only + approved scan commands), performance profiler (benchmarks in isolated worktree), accessibility checker (a11y tools + WCAG 2.1 checks), docs writer (doc paths only), and verifier (approved verification commands only).\n\nFeatures include parallel audit phase (4 review agents run concurrently), TDD mode (hook-level test-first enforcement), custom roles via manifest (--custom-role), REVIEW.md injection for project-specific review criteria, mission timeout and failure tracking.\n\n96 tests. 100% Python stdlib. Zero external dependencies.","developer":"AgentBuildersApp","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/AgentBuildersApp/eight-eyes.git"},{"slug":"dopamine","title":"dopamine","description":"Dopamine is a Claude Code hook plugin with two modes. In Degrade mode, it\nopens YouTube Shorts or Instagram Reels in a compact window next to your\nterminal while Claude handles your tasks — and auto-closes when the work is\ndone. In Progress mode, it detects what Claude is currently working on (React,\nPython, Docker, etc.) and opens a relevant best-practices article that stays\nopen for reading. Switch between guilt-free procrastination and passive\nlearning with a single command.","developer":"Teese999","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/Teese999/Dopamine.git"},{"slug":"domain-check-skill","title":"domain-check-skill","description":"Ask your AI coding agent if a domain is free — and get a real answer. Works for any TLD, including .de / .eu / .at / .ch where most checkers give up. Plugin for Claude Code. Pure Python, zero setup.","developer":"PleasePrompto","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/PleasePrompto/domain-check-skill.git"},{"slug":"docvet-lsp","title":"docvet-lsp","description":"Real-time docstring quality diagnostics for Python files","developer":"Alberto-Codes","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/Alberto-Codes/docvet.git"},{"slug":"docvet","title":"docvet","description":"Real-time Python docstring quality diagnostics via LSP. Detects missing Raises, Yields, Attributes, Examples, and cross-reference sections in Google-style docstrings. Also checks for griffe rendering warnings and missing __init__.py files for mkdocs discoverability. Findings appear inline as you code — required rules as warnings, recommended rules as hints.","developer":"Alberto-Codes","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/Alberto-Codes/docvet.git"},{"slug":"dishonest-code-audit","title":"dishonest-code-audit","description":"Find code that lies to users: silent failures plus mock/stub/placeholder code in production paths. Supports JS/TS, Python, Go, Rust, Ruby via pluggable stack profiles.","developer":"Yonatan Hyatt","category":null,"reported_installs":null,"active":true,"first_seen_at":"2026-10-03T22:13:33.942Z","last_seen_at":"2026-11-08T00:00:01.297Z","last_changed_at":"2026-10-03T22:13:33.942Z","platform":"claude","sources":["Community"],"works_with":["Claude Code"],"marketplace_url":null,"repository_url":"https://github.com/yhyatt/dishonest-code-audit.git"}],"total":112,"page":2,"per_page":36,"observed_at":"2026-11-08T00:00:01.297Z","source_counts":{"web":529,"demo":12,"skills":5,"official":317,"community":2283,"knowledge-work":123}}