Daily AI Implementation Scout Council

Today's ranked top 20

#1

newbuild nowBoth runtimes32 / 35

microsoft/SkillOpt repo

Text-space optimizer from Microsoft Research that trains reusable natural-language skills for frozen LLM agents through trajectory-driven edits and validation-gated updates. Produces deployable best_skill.md artifacts. Ships SkillOpt-Sleep in v0.2.0: a nightly self-evolution engine that harvests real agent session data, mines improvement opportunities, replays and consolidates them behind a held-out validation gate. Claude Code integration shell is in the repo.

What it does for you: AXION updates skills manually. SkillOpt replaces that loop with a research-backed automated pipeline: SkillOpt-Sleep reads real Claude Code session trajectories, identifies skill weaknesses, proposes edits, validates them against held-out tests, and writes an improved best_skill.md. The alirezarezvani/claude-skills collection (rank 2) already vendors a SkillOpt-Sleep integration, which is the clearest community signal that it is production-grade. This is the first time this repo has appeared in the registry.

In practice: 16,100 GitHub stars and 1,500 forks as of 2026-08-19. Microsoft Research blog feature on 2026-07-24 with coverage from VentureBeat, Synced, and The Decoder. The community validation comes from the alirezarezvani/claude-skills pack (24,600 stars) having already vendored the SkillOpt-Sleep shell.

For: Both runtimes. Python package (pip install skillopt). Ships integration shells for Claude Code, Codex, Copilot, and Devin in the repo. SKILL.md artifacts are format-agnostic and run in Hermes or any skill-compatible harness.

Security4

Quality5

Auditability5

Useful to you5

Useful to community5

Buildable now4

Hermes4

Verdict: build now. 32 total. pip install skillopt and the Claude Code integration shell in the repo makes this a same-day install. The nightly self-evolution loop directly addresses the AXION skill-quality maintenance problem. No equivalent validation-gated, trajectory-driven skill trainer exists in the registry.

Build #1 microsoft/SkillOpt: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/microsoft/SkillOpt. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · 16,100 stars, 1,500 forks as of 2026-08-19. v0.2.0 released 2026-07-02 on PyPI. Microsoft Research blog post 2026-07-24. VentureBeat and The Decoder coverage same week. MIT license.

#2

build nowBoth runtimes32 / 35

alirezarezvani/claude-skills skill

Community collection of 345 Claude Code and agent skills across 10 domains: engineering core (52 skills), engineering powerful (84 skills including agent-harness, zero-hallucination-coder, and skillopt-sleep), product (17), marketing, compliance, C-level advisory, research, business operations, commercial and finance, and daily productivity. Ships 30+ prebuilt agents and 70+ custom commands.

What it does for you: YY can clone this repo and cherry-pick from 345 battle-tested skills rather than writing from scratch. The agent-harness skill adds a goal-plan-execute-verify-close loop over any domain. The engineering/skillopt-sleep module (rank 1 today) ties this collection to automated nightly skill improvement. Updated 2026-08-18, one day before today's issue. Last listed 2026-08-09 as watch; upgraded here to build now on re-review after the skillopt-sleep integration and 10-day update cycle were confirmed.

In practice: 24,600 GitHub stars and 3,500 forks as of 2026-08-19. Updated 2026-08-18. The breadth at 345 skills signals a systematic build across a full workday. The domain table in the README maps each domain to highlights and file paths, which is unusual rigor for a community skill repo.

For: Both runtimes. Markdown skill files with no build step; compatible with Claude Code, Codex, Hermes, and any skill-runner that reads SKILL.md or CLAUDE.md format.

Security4

Quality4

Auditability4

Useful to you5

Useful to community5

Buildable now5

Hermes5

Verdict: build now. 32 total. git clone and copy individual skills today. Two material changes since last listing (2026-08-09): the skillopt-sleep integration was added and the engineering/powerful tier expanded. Last verdict was watch; upgraded to build now on re-review.

Build #2 alirezarezvani/claude-skills: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/alirezarezvani/claude-skills. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · 24,600 GitHub stars, 3,500 forks as of 2026-08-19. Updated 2026-08-18. 345 skills across 10 domains. Last seen in registry 2026-08-09 with verdict watch.

#3

newtest firstClaude (Axion)30 / 35

diet103/claude-code-infrastructure-showcase repo

Reference implementation of Claude Code infrastructure patterns: a skill-activation-prompt hook that fires on every user prompt and checks a skill-rules.json file for trigger patterns, suggesting relevant skills automatically. Described as the solution to the number-one problem with Claude Code skills: they do not self-activate. Ships a hooks README, agents README, skills README, and a Claude integration guide written for AI-assisted setup.

What it does for you: AXION already has hooks in settings.json but no systematic trigger-based skill activation. Copying the two essential hooks and one skill-rules.json from this repo introduces declarative, auditable auto-activation. An afternoon of work connects this to the existing AXION skill library.

In practice: 10,000 GitHub stars as of 2026-08-19. The README leads with a practical quick-start. The Claude Integration Guide is written specifically for AI-assisted setup, which is the mode AXION operates in.

For: Claude (Axion). Claude Code-specific: the skill-activation-prompt hook and skill-rules.json pattern are designed for Claude Code's hook and skill runtime. Hermes does not run the same hook lifecycle.

Security4

Quality4

Auditability4

Useful to you5

Useful to community4

Buildable now5

Hermes4

Verdict: test first. 30 total. The auto-activation pattern is immediately applicable to AXION. Test on an isolated project first to confirm the hook does not conflict with the existing build-gate hook already in AXION settings.json.

Build #3 diet103/claude-code-infrastructure-showcase: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/diet103/claude-code-infrastructure-showcase. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · 10,000 GitHub stars as of 2026-08-19. Updated 2026-08-19. TypeScript, MIT license. Ships CLAUDE_INTEGRATION_GUIDE.md plus three README files covering skills, hooks, and agents.

#4

newtest firstBoth runtimes29 / 35

akitaonrails/ai-memory repo

Rust CLI for long-term memory for agent coding sessions and cross-vendor handoff. Stores structured task context, failed approaches, open questions, and architecture notes in a vendor-agnostic format. When you switch from Claude Code to OpenAI Codex in the same directory, the next agent reads the memory and continues without re-explaining the project. Ships a web UI for memory inspection and proactive query routing via Agent Skills.

What it does for you: AXION sessions switch between Claude Code and Hermes; context is currently lost unless YY writes handoffs manually. ai-memory automates that handoff: after any Claude Code session the memory updates, and the next session (Claude, Codex, or Hermes) picks it up automatically. The Agent Skills integration means routing guidance travels with the memory file, matching AXION's existing skill format.

In practice: 2,673 GitHub stars and 231 forks as of 2026-08-19, with 730 new stars today: the number-one trending Rust repo on GitHub today. Rust 1.95+, MIT license. README covers handoff examples, proactive query routing, and raw-wiki inspection commands.

For: Both runtimes. Rust CLI binary that writes a vendor-agnostic memory block and Agent Skills files into any project directory. Works with Claude Code, Codex, Gemini CLI, and any agent CLI. Hermes can read the resulting memory files at session startup.

Security4

Quality4

Auditability4

Useful to you5

Useful to community5

Buildable now4

Hermes3

Verdict: test first. 29 total. Directly solves AXION's cross-session memory loss. Rust binary avoids Python dependency conflicts. Test on one non-critical project first to confirm the memory format does not expose sensitive AXION architecture in a shared directory.

Build #4 akitaonrails/ai-memory: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/akitaonrails/ai-memory. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · 2,673 stars, 231 forks, 730 new stars on 2026-08-19 (number-one trending Rust repo today). Rust 1.95+, MIT license. Release badge visible in README.

#5

test firstClaude (Axion)29 / 35

modelcontextprotocol/servers plugin

Anthropic-maintained collection of reference and community MCP server implementations: filesystem, GitHub, Slack, Google Drive, Postgres, fetch, and others. The canonical starting point for adding MCP tools to Claude Code. Re-listed after 8 weeks (last seen 2026-06-24).

What it does for you: AXION uses Claude Code with MCP tools daily. The server list has grown since June; the GitHub and fetch servers are directly relevant to AXION pipelines and can be installed today without any build step.

In practice: Continuously maintained by Anthropic and the MCP community. Re-listed here as the 8-week reminder that several servers added since June are worth evaluating for AXION.

For: Claude (Axion). Official MCP server reference implementations; each server is installed independently into Claude Code or any MCP client via the mcpServers block in settings.json.

Security4

Quality5

Auditability5

Useful to you3

Useful to community5

Buildable now4

Hermes3

Verdict: test first. 29 total. Re-listed after 8 weeks. The GitHub and fetch servers score buildable_now:5 individually; the collection as a whole scores 4 because individual server selection and config is required.

Install: enable individual servers by adding them to the mcpServers block in ~/.claude/settings.json following each server's README. Review before enabling. Copy

Source · Official Anthropic repository. MCP org has 24,000+ GitHub followers. Continuously updated; last commit within 7 days per the repo page as of scouting. Last seen in registry 2026-06-24.

#6

newtest firstClaude (Axion)28 / 35

context7 (claude-plugins-official external plugin) plugin

MCP server that injects up-to-date library documentation into Claude Code context. When code references a library, context7 fetches current docs for that library version and adds them as context, preventing hallucinated API calls based on outdated training data. Listed in the Anthropic official external plugins directory.

What it does for you: AXION skills and agents frequently call Python library APIs that change between versions. context7 eliminates the class of errors where Claude generates code for an outdated API signature. Anthropic has reviewed this integration for listing in the official directory.

In practice: Listed in the Anthropic-managed external_plugins directory alongside GitHub, GitLab, Discord, Asana, and Firebase. Widely referenced in the Claude Code community as a standard quality-of-life install. Not currently enabled in YY's settings.

For: Claude (Axion). MCP server installed as a Claude Code external plugin via anthropics/claude-plugins-official. Not applicable to Hermes directly.

Security4

Quality4

Auditability4

Useful to you4

Useful to community4

Buildable now5

Hermes3

Verdict: test first. 28 total. Directly installable via the official plugin directory. Not currently in YY's enabledPlugins. The documentation accuracy improvement is measurable immediately by comparing generated API calls against the current library docs.

Install: add context7 to enabledPlugins in ~/.claude/settings.json following the external_plugins/context7 README in anthropics/claude-plugins-official. Not currently in YY's enabledPlugins. Review before enabling. Copy

Source · Listed at github.com/anthropics/claude-plugins-official/tree/main/external_plugins/context7 as of 2026-08-19. Anthropic-reviewed. 6 external plugins total in the directory: asana, context7, discord, fakechat, firebase, github, gitlab.

#7

newtest firstClaude (Axion)28 / 35

github (claude-plugins-official external plugin) plugin

Official GitHub MCP server for Claude Code. Provides direct GitHub API access from within a Claude Code session: read repos, files, issues, and PRs; create branches; commit files; open PRs. Listed in the Anthropic official external plugins directory.

What it does for you: AXION runs git operations manually via Bash tools today. The GitHub plugin lets Claude Code read issues, create branches, and open PRs directly, reducing tool-call overhead for the AXION docs repo and future tracked work. Anthropic has reviewed this integration.

In practice: Listed in the Anthropic-managed external_plugins directory. GitHub integration is the most-requested MCP server type across the community. Not currently enabled in YY's settings.

For: Claude (Axion). MCP server installed as a Claude Code external plugin via anthropics/claude-plugins-official. Not applicable to Hermes directly.

Security3

Quality4

Auditability4

Useful to you4

Useful to community5

Buildable now5

Hermes3

Verdict: test first. 28 total. Direct GitHub access from Claude Code is high-value. Not currently in YY's enabledPlugins. Test on a non-critical repo first to confirm write behavior before enabling on the main AXION docs repo.

Install: add the github external plugin to enabledPlugins in ~/.claude/settings.json following the external_plugins/github README in anthropics/claude-plugins-official. Not currently in YY's enabledPlugins. Review before enabling. Copy

Source · Listed at github.com/anthropics/claude-plugins-official/tree/main/external_plugins/github as of 2026-08-19. Anthropic-reviewed external plugin.

#8

watchBoth runtimes28 / 35

BerriAI/litellm repo

AI gateway with Rust core and Python SDK. Routes calls to 100+ LLM providers through a single OpenAI-format endpoint. Adds cost tracking, guardrails, load balancing, and logging. Supports A2A agent invocation for LangGraph, Vertex AI, Bedrock AgentCore, and Pydantic AI agents. YC W23. Ships Terraform modules for AWS and GCP.

What it does for you: AXION calls Claude directly with no cost tracking across sessions. Adding LiteLLM as a local proxy gives YY a unified cost view, the ability to swap providers without code changes, and a logging layer. The A2A support means AXION agents can call external framework agents through the same gateway.

In practice: 10,700 GitHub forks as of 2026-08-19, signalling heavy production use. YC W23. Rust core adds minimal latency. The docs have moved to a separate repo (BerriAI/litellm-docs), a maturity signal.

For: Both runtimes. Python SDK plus a self-hosted proxy gateway. Agents in both Claude Code and Hermes call it via a standard OpenAI-format endpoint without changing their calling code.

Security3

Quality4

Auditability4

Useful to you4

Useful to community5

Buildable now4

Hermes4

Verdict: watch. 28 total. High-quality but held at watch because it introduces a proxy layer between AXION and the Anthropic API, which is a significant infrastructure commitment. Score justifies test first but the setup cost and operational dependency push verdict to watch.

Build #8 BerriAI/litellm: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/BerriAI/litellm. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · 10,700 GitHub forks as of 2026-08-19. YC W23. 100+ LLM providers documented at docs.litellm.ai. MIT license. Terraform module at registry.terraform.io/modules/BerriAI/litellm.

#9

watchBoth runtimes28 / 35

instructor-ai/instructor repo

Python library for structured output extraction from LLMs using Pydantic models. Adds retry logic, validation, and streaming to raw LLM calls. Supports Anthropic, OpenAI, Google, and others with a unified interface. Last seen in registry 2026-06-23 with verdict build now.

What it does for you: AXION agents frequently parse LLM output into structured data. Instructor replaces ad-hoc JSON parsing and retry logic with a typed, validated, and retry-aware interface. The Anthropic provider integration works with existing Claude calls without a provider switch.

In practice: Stable and widely adopted. No dramatic new release since June but the tool has compounded into a standard in the structured-output space. Re-listed as an 8-week reminder that AXION has not yet adopted it.

For: Both runtimes. Python package; provider-agnostic structured output extraction. Hermes can use it directly for typed response validation from any LLM call.

Security4

Quality4

Auditability4

Useful to you4

Useful to community5

Buildable now3

Hermes4

Verdict: watch. 28 total. Scores justify test first but held at watch because urgency has settled since the 2026-06-23 build-now listing and AXION has not yet prioritised structured output standardisation. The score discrepancy is intentional and noted in dissent.

Build #9 instructor-ai/instructor: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/instructor-ai/instructor. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · First seen in registry 2026-06-23 with verdict build now. pip install instructor. MIT license. Anthropic provider supported directly. Thousands of GitHub stars per prior issues.

#10

watchBoth runtimes27 / 35

openai/openai-agents-python repo

OpenAI's lightweight multi-agent workflow SDK for Python. Provides agents, handoffs, guardrails, and tracing primitives. Updated today (2026-08-19). Last seen in registry 2026-08-14 as watch.

What it does for you: YY designs multi-agent workflows for AXION. The handoff and guardrail patterns here are directly comparable to AXION's current design. The tracing integration ships structured trace output that could inform AXION agent logging. Updated today while scouting was running.

In practice: 28,800 GitHub stars as of 2026-08-19. Updated within the past hour of scouting today. OpenAI official SDK.

For: Both runtimes. Python package; handoff and guardrail patterns apply to any agent framework including Hermes regardless of the underlying LLM provider.

Security3

Quality5

Auditability4

Useful to you3

Useful to community5

Buildable now4

Hermes3

Verdict: watch. 27 total. Re-listed because it updated today. Useful as a pattern reference but the OpenAI provider coupling limits direct adoption in AXION's Claude-first stack.

Build #10 openai/openai-agents-python: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/openai/openai-agents-python. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · 28,800 GitHub stars as of 2026-08-19. Updated 2026-08-19. Last seen in registry 2026-08-14.

#11

newwatchStandalone tool26 / 35

bojieli/ai-agent-book repo

Open-source book titled Deep Understanding of AI Agents: Design Principles and Engineering Practice, by Li Bojie. Full text in Chinese with community translations to English, Spanish, Indonesian, and Traditional Chinese. Covers agent design principles and engineering practices with companion code by chapter. Available as PDF and EPUB from the Releases page.

What it does for you: YY makes AXION architectural decisions daily. This book is the most comprehensive free reference on agent engineering available in August 2026, covering the same design trade-offs that appear in AXION build decisions. The English translation makes it accessible. Trending at 556 new stars today.

In practice: 39,100 GitHub stars and 4,300 forks as of 2026-08-19. 556 new stars today, the highest single-day gain of any book repo in the agent space this week. Community translations to 5 languages signal broad readership. The author is a researcher at a major Chinese institution.

For: Standalone tool. Open-source book with companion code. The PDF and EPUB are standalone reading material; the chapter code examples run independently but do not integrate into Claude Code or Hermes.

Security5

Quality4

Auditability5

Useful to you4

Useful to community5

Buildable now2

Hermes1

Verdict: watch. 26 total. Excellent reference but not installable infrastructure. Scores low on buildable_now and hermes_usable because it produces no artifact for AXION. Worth reading before the next AXION agent architecture decision.

Build #11 bojieli/ai-agent-book: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/bojieli/ai-agent-book. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · 39,100 GitHub stars, 4,300 forks as of 2026-08-19. 556 new stars on 2026-08-19. PDF and EPUB at Releases. Community translations in 5 languages.

#12

watchBoth runtimes26 / 35

lastmile-ai/mcp-agent repo

MCP-native Python agent framework built on the Anthropic Building Effective Agents patterns. Composable workflow patterns, Temporal-backed durability, structured logging, and token accounting. Any MCP server connects without custom adapters. Last seen in registry 2026-07-13 with verdict watch.

What it does for you: AXION already runs MCP servers. mcp-agent lets YY compose those MCP connections into typed workflow patterns without writing custom orchestration. The Temporal integration at the framework level means durable long-running workflows come for free once adopted.

In practice: 8,500 GitHub stars per prior scouting. Re-listed because 5+ weeks have passed and it remains the only MCP-first agent framework in the registry. However, the GitHub search result today listed the last update as January 2026, which is 7 months ago; verify active maintenance before adopting.

For: Both runtimes. Python package that wraps MCP servers into composable agent workflow patterns. Hermes can call mcp-agent workflows as standard Python functions.

Security3

Quality4

Auditability4

Useful to you4

Useful to community4

Buildable now3

Hermes4

Verdict: watch. 26 total. Re-listed after 5+ weeks. The 7-month stale update date reduces the buildable_now score. Verify active maintenance before adopting; if the January 2026 date is confirmed, this moves toward skip at next review.

Build #12 lastmile-ai/mcp-agent: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/lastmile-ai/mcp-agent. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · 8,500 GitHub stars per prior scouting. Last GitHub update listed as January 2026 in today's search results. Python, Apache license. docs.mcp-agent.com. Last seen in registry 2026-07-13.

#13

newwatchBoth runtimes25 / 35

DougTrajano/pydantic-ai-skills skill

Python package that adds Agent Skills support with progressive disclosure to Pydantic AI. Supports filesystem-loaded skills (from SKILL.md files) and programmatic skills defined in code. Bridges the agentskills.io format with Pydantic AI dependency injection and agent lifecycle. Updated 8 hours before scouting.

What it does for you: AXION uses the SKILL.md format throughout. If Pydantic AI agents are added to AXION (a growing pattern in the ecosystem), this package means the existing AXION skill library works without format conversion.

In practice: 359 GitHub stars as of 2026-08-19. Updated 8 hours ago. Small but focused. The acknowledgments credit Anthropic for the Agent Skills format and the Pydantic AI team for the framework, signalling a practitioner who read the specs before building.

For: Both runtimes. Python package implementing the Agent Skills format for Pydantic AI. Resulting skills are SKILL.md-compatible and loadable by Hermes or any skill-aware runner.

Security4

Quality3

Auditability3

Useful to you4

Useful to community3

Buildable now4

Hermes4

Verdict: watch. 25 total. Small project (359 stars) but fresh and directly aligned with AXION's skill format. Watch for 30 days to see if Pydantic AI adoption in AXION's ecosystem grows before integrating.

Build #13 DougTrajano/pydantic-ai-skills: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/DougTrajano/pydantic-ai-skills. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · 359 GitHub stars as of 2026-08-19. Updated 2026-08-19, 8 hours before scouting. Python package, MIT license. Documentation at dougtrajano.github.io/pydantic-ai-skills.

#14

watchBoth runtimes25 / 35

temporalio/temporal repo

Durable workflow execution engine. Runs long-running agent tasks as code that survives failures, retries, and restarts automatically. Used by teams running production multi-agent workflows that must not lose state on crash. Last seen in registry 2026-06-30 with verdict test first.

What it does for you: AXION nightly routines (ingest, newsletter, scout) lose state if they crash mid-run. Temporal makes those workflows resumable from the last checkpoint without re-running from the start. The Python SDK is the relevant integration path.

In practice: Production infrastructure used by Coinbase, Stripe, and Netflix at scale. Re-listed after 7 weeks as a standing reminder that AXION's nightly routines are still not crash-durable.

For: Both runtimes. Durable workflow engine with Python, Go, Java, and TypeScript SDKs. AXION agents call Temporal workflows from Python; Hermes can submit work to a Temporal cluster.

Security4

Quality4

Auditability4

Useful to you4

Useful to community4

Buildable now2

Hermes3

Verdict: watch. 25 total. Held at watch because running a Temporal server adds operational overhead. Test first verdict from 2026-06-30 stands in principle; held at watch here due to setup cost.

Build #14 temporalio/temporal: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/temporalio/temporal. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · Python SDK at pypi.org/project/temporalio. Production case studies for Coinbase, Stripe, and Netflix at temporal.io. Last seen in registry 2026-06-30.

#15

watchBoth runtimes24 / 35

opendatalab/MinerU repo

Document-to-LLM-markdown converter. Transforms PDFs, Word, and Office documents into clean markdown or JSON for agentic workflows. Latest release is v3.4.5 as of 2026-08-19. Last seen in registry 2026-07-01 with verdict test first.

What it does for you: AXION brain ingest regularly pulls in PDFs and documents. MinerU converts them to markdown without manual intervention, giving the ingest pipeline structured text the LLM can index and cross-reference. v3.4.5 is a material version increment since last listing.

In practice: Re-listed after 7+ weeks. v3.4.5 is the current latest release per the Releases page visited today. The tool is recommended across AI workflow threads as the best open-source PDF-to-markdown converter.

For: Both runtimes. Python package; Claude Code and Hermes agents call it as a pre-processing step in any document ingestion pipeline.

Security3

Quality4

Auditability4

Useful to you3

Useful to community4

Buildable now3

Hermes3

Verdict: watch. 24 total. Re-listed because v3.4.5 represents a material update since the 2026-07-01 listing. Held at watch because PDF ingestion for the AXION brain has not been prioritised.

Build #15 opendatalab/MinerU: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/opendatalab/MinerU. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · Latest release v3.4.5 as of 2026-08-19, confirmed from the Releases page. Last seen in registry 2026-07-01. Python, Apache license. Documentation at opendatalab.github.io/MinerU.

#16

newwatchClaude (Axion)23 / 35

chaitanyagiri/munder-difflin repo

Local multi-agent harness in TypeScript and Electron. Spawns multiple agent processes, connects them through a shared hive: a local git repo of plain files with outbox and inbox routing. A GOD agent orchestrator resolves routine tasks autonomously and escalates only spend decisions, destructive operations, and scope changes to a human approval queue. Ships HIVE.md and SPEC.md as architecture documents.

What it does for you: The single-committer git design solves the index.lock corruption problem that affects AXION multi-agent runs. The GOD agent pattern maps to AXION's orchestrator-then-workers hierarchy. Study HIVE.md and SPEC.md to extract the pattern for AXION without adopting the Electron runtime.

In practice: 1,974 GitHub stars and 236 forks as of 2026-08-19, with 256 new stars today (trending TypeScript). The name is comedic but HIVE.md is a serious technical spec. TypeScript, MIT.

For: Claude (Axion). Electron desktop app that spawns claude, agy, codex, or custom CLI agents as terminal PTY processes. Claude Code is a natively supported provider. Hermes is not supported but could be added as a custom process.

Security3

Quality4

Auditability4

Useful to you4

Useful to community3

Buildable now3

Hermes2

Verdict: watch. 23 total. The hive architecture and GOD agent pattern are worth studying. The Electron runtime is a mismatch for AXION's headless CLI and bridge architecture, so direct adoption is unlikely without extracting the pattern as a standalone library.

Build #16 chaitanyagiri/munder-difflin: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/chaitanyagiri/munder-difflin. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · 1,974 GitHub stars, 236 forks, 256 new stars on 2026-08-19 (trending TypeScript today). TypeScript, MIT. Ships HIVE.md and SPEC.md.

#17

watchClaude (Axion)23 / 35

modelcontextprotocol/registry plugin

Community-driven registry for MCP servers, exposing a REST API for server discovery and metadata. Published as a Docker image. The live API is at registry.modelcontextprotocol.io. Last seen in registry 2026-07-15.

What it does for you: When YY needs a new MCP server for AXION, querying the registry is faster than a manual GitHub search. The live API returns structured server metadata including description, install instructions, and version.

In practice: Re-listed after 5+ weeks. The registry has grown since July with new MCP server submissions. Re-listed as a standing awareness item.

For: Claude (Axion). Community MCP registry service exposing a REST API for server discovery. Claude Code uses it for discovery; AXION automation scripts can query it directly.

Security4

Quality3

Auditability3

Useful to you3

Useful to community4

Buildable now3

Hermes3

Verdict: watch. 23 total. Useful reference, not a buildable item for AXION. Re-listed as a reminder to use the registry before searching GitHub manually for MCP servers.

Install: query registry.modelcontextprotocol.io/api/v0/servers for server discovery; install individual servers separately per their READMEs. Review before enabling. Copy

Source · Live API at registry.modelcontextprotocol.io. Docker image at ghcr.io/modelcontextprotocol/registry with tagged releases (v1.0.0, v1.1.0, etc.) per the README. Last seen in registry 2026-07-15.

#18

newwatchClaude (Axion)23 / 35

typescript-lsp (claude-plugins-official plugin) plugin

Anthropic-official Claude Code plugin that connects a TypeScript Language Server Protocol instance to Claude Code. Gives Claude Code real-time TypeScript type-checking, go-to-definition, and refactoring support within its editing context. Listed alongside skill-creator and swift-lsp in the official plugins folder.

What it does for you: AXION includes TypeScript projects (the Telegram bridge, the scout site, and several TOOL-AND-APP tools). The typescript-lsp plugin lets Claude Code resolve types and imports accurately when editing those projects, reducing type-error hallucinations. Not currently enabled in YY's settings.

In practice: Anthropic-built official plugin. The official provenance means Anthropic has tested and maintains this integration.

For: Claude (Axion). Anthropic-official Claude Code plugin connecting a TypeScript LSP instance to Claude Code sessions. Not applicable to Hermes.

Security4

Quality4

Auditability4

Useful to you3

Useful to community3

Buildable now4

Hermes1

Verdict: watch. 23 total. Useful for TypeScript work but AXION's TypeScript projects are not currently the primary development surface. Watch until a TypeScript project is actively in development.

Install: add typescript-lsp to enabledPlugins in ~/.claude/settings.json following the plugins/typescript-lsp README in anthropics/claude-plugins-official. Not currently in YY's enabledPlugins. Review before enabling. Copy

Source · Listed at github.com/anthropics/claude-plugins-official/tree/main/plugins/typescript-lsp as of 2026-08-19. Anthropic-built official plugin, co-listed with skill-creator and swift-lsp.

#19

watchBoth runtimes22 / 35

volcengine/OpenViking repo

Self-evolving context database for AI agents that unifies agent memory, knowledge RAG, and skills in a single queryable store. Last seen in registry 2026-07-13 with verdict build now.

What it does for you: AXION manages memory (MEMORY.md), knowledge (YY-2nd Brain), and skills (AXION/Skills) as three separate systems. OpenViking unifies these into one context database with a self-evolution layer. Re-listed after 5+ weeks because AXION's memory architecture question recurs across sessions.

In practice: Re-listed after 5+ weeks (last 2026-07-13). Was build now at first listing; held at watch here because unified context database migration is a larger architectural commitment than initially estimated.

For: Both runtimes. Python and TypeScript SDKs; the context database runs as a service that AXION or Hermes agents call via HTTP or the SDK.

Security3

Quality3

Auditability3

Useful to you4

Useful to community3

Buildable now3

Hermes3

Verdict: watch. 22 total. Downgraded from build now (2026-07-13) because integrating a unified context database requires migrating existing AXION memory and brain structures, which is a high-risk multi-session effort.

Build #19 volcengine/OpenViking: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/volcengine/OpenViking. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · Last seen in registry 2026-07-06 and 2026-07-13 with verdict build now. openviking.ai documentation site. Python and TypeScript SDKs per the GitHub repo.

#20

newwatchStandalone tool18 / 35

genlayerlabs/genlayer-project-boilerplate repo

TypeScript boilerplate for building AI agent contracts on GenLayer, a decentralized platform where LLMs act as validators and executors for smart contracts. Trending at 543 new stars today, the 5th highest trending repo overall on GitHub.

What it does for you: Provides context on the emerging decentralized AI agent space. Not applicable to AXION today but the pattern of LLMs as on-chain validators is a design direction worth monitoring for MCL platform strategy.

In practice: 15,893 GitHub stars, 800 forks, 543 new stars today. TypeScript, MIT. Novel: AI agents as blockchain validators rather than off-chain assistants.

For: Standalone tool. TypeScript boilerplate for GenLayer, a decentralized AI agent platform with its own blockchain-backed runtime. Not integrable with Claude Code or Hermes directly.

Security3

Quality3

Auditability3

Useful to you2

Useful to community4

Buildable now2

Hermes1

Verdict: watch. 18 total. Below the 20-point watch floor but included because it is the 5th overall trending repo today and deserves a registry record. If no AXION relevance emerges in 30 days it moves to skip.

Build #20 genlayerlabs/genlayer-project-boilerplate: use the ai-implementation-build-intake skill to build this safely. Source: https://github.com/genlayerlabs/genlayer-project-boilerplate. Save canonical skill/agent under AXION\Skills and AXION\Agents. Copy

Source · 15,893 GitHub stars, 800 forks, 543 new stars on 2026-08-19 (5th overall trending repo on GitHub today). TypeScript, MIT license.