Morning Singularity Digest - 2026-06-23

Context

The agent harness performance optimization system.

What's new

Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

Key details

• Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
• Language: English | Português (Brasil) | 简体中文 | 繁體中文 | 日本語 | 한국어 | Türkçe | Русский | Tiếng Việt | ไทย | Deutsch | Español Warning Official sources only.
• Install ECC only from verified channels: the GitHub repository github.com/affaan-m/ECC, the npm packages ecc-universal and ecc-agentshield, the GitHub App, the plugin slug ecc@ecc, and the project website ecc.tools.
• Third-party re-uploads and unofficial mirrors are not maintained or reviewed by the project and may contain malware.

Results & evidence

• 211.9K+ stars | 32.5K+ forks | 230+ contributors | 12+ language ecosystems | Cross-harness agent workflows Language / 语言 / 語言 / Dil / Язык / Ngôn ngữ / Idioma English | Português (Brasil) | 简体中文 | 繁體中文 | 日本語 | 한국어 | Türkçe | Русский | Tiếng Việt | ไทย | Deu...
• Production-ready agents, skills, hooks, rules, MCP configurations, and legacy command shims evolved over 10+ months of intensive daily use building real products.
• ECC v2.0.0 adds the public Hermes operator story on top of that reusable layer: start with the Hermes setup guide, then review the 2.0.0 release notes and cross-harness architecture.

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

The open-source app everyone uses to manage agents at work Quickstart · Docs · GitHub · Discord · Twitter · Website full-tour.webm Open-source orchestration for teams of AI agents.

What's new

The open-source app everyone uses to manage agents at work Quickstart · Docs · GitHub · Discord · Twitter · Website full-tour.webm Open-source orchestration for teams of AI agents.

Key details

• If OpenClaw is an employee, Paperclip is the company.
• Paperclip is a Node.js server and React UI that orchestrates a team of AI agents to run a business.
• Bring your own agents, assign goals, and track work and costs from one dashboard.
• Under the hood: org charts, budgets, governance, goal alignment, and agent coordination.

Results & evidence

• | Step | Example | | |---|---|---| | 01 | Define the goal | "Build the #1 AI note-taking app to $1M MRR." | | 02 | Hire the team | CEO, CTO, engineers, designers, marketers — any bot, any provider.
• | | 03 | Approve and run | Review strategy.
• | - ✅ You want to build autonomous AI companies - ✅ You coordinate many different agents (OpenClaw, Codex, Claude, Cursor) toward a common goal - ✅ You have 20 simultaneous Claude Code terminals open and lose track of what everyone is doing - ✅ You want age...

Limitations / unknowns

• When they hit the limit, they stop.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

A persistent challenge for nugget-based evaluation is the need to manually curate sets of nuggets for each topic in a test collection -- a laborious process that scales poorly to novel information needs.

What's new

arXiv:2605.04458v2 Announce Type: replace Abstract: Evaluation of long-form, citation-backed reports has lately received significant attention due to the wide-scale adoption of retrieval-augmented generation (RAG) systems.

Key details

• Core to many evaluation frameworks is the use of atomic facts, or nuggets, to assess a report's coverage of query-relevant information attested in the underlying collection.
• While nuggets have traditionally been represented as short statements, recent work has used question-answer (QA) representations, enabling fine-grained evaluations that decouple the information need (i.e.
• the question) from the potentially diverse content that satisfies it (i.e.
• A persistent challenge for nugget-based evaluation is the need to manually curate sets of nuggets for each topic in a test collection -- a laborious process that scales poorly to novel information needs.

Results & evidence

• arXiv:2605.04458v2 Announce Type: replace Abstract: Evaluation of long-form, citation-backed reports has lately received significant attention due to the wide-scale adoption of retrieval-augmented generation (RAG) systems.
• Accordingly, we introduce DoGMaTiQ, a pipeline for generating high-quality QA-based nugget sets in three stages: (1) document-grounded nugget generation, (2) paraphrase clustering, and (3) nugget subselection based on principled quality criteria.
• Computer Science > Computation and Language [Submitted on 6 May 2026 (v1), last revised 19 Jun 2026 (this version, v2)] Title:DoGMaTiQ: Automated Generation of Question-and-Answer Nuggets for Report Evaluation View PDFAbstract:Evaluation of long-form, citat...

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

To address these challenges, we propose CodeTeam, an LLM-based multi-agent framework that separates planning, decision making, and implementation into distinct, coordinated stages.

What's new

To address these challenges, we propose CodeTeam, an LLM-based multi-agent framework that separates planning, decision making, and implementation into distinct, coordinated stages.

Key details

• Compared with function-level code generation, this task demands longer planning horizons, stable interfaces across files, and iterative debugging of cross-file inconsistencies.
• To address these challenges, we propose CodeTeam, an LLM-based multi-agent framework that separates planning, decision making, and implementation into distinct, coordinated stages.
• In the planning stage, multiple Architect agents draft competing software design sketches (SDS), optionally grounded by retrieved design references.
• A CTO agent then evaluates, selects, and normalizes the most promising SDS into a machine-checkable contract that specifies file ownership, public interfaces, and dependency constraints.

Results & evidence

• arXiv:2606.22082v1 Announce Type: cross Abstract: Natural language to repository generation (NL2Repo) requires a system to construct an entire software repository from a natural-language requirements document.
• On the synthesis-based SketchEval benchmark, we explicitly compare CodeTeam's prompt-engineering (PE) and supervised fine-tuning (SFT) variants with the corresponding CodeS variants, where CodeTeam improves the overall SketchBLEU by 4.1 and 2.9 absolute poi...
• On the execution-based NL2Repo-Bench benchmark, used as an external validation protocol, CodeTeam achieves the highest average test pass rate in both settings (34.6% PE, 42.3% SFT), confirming that the sketch-improvements extend to functional correctness un...

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

Show HN: Route LLM prompts to cheapest capable model – pydantic-AI and litellm

What's new

Show HN: Route LLM prompts to cheapest capable model – pydantic-AI and litellm

Key details

• Show HN: Route LLM prompts to cheapest capable model – pydantic-AI and litellm

Results & evidence

• No hard numbers surfaced in the source text; treat claims as directional until benchmarks appear.

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

The open-source app everyone uses to manage agents at work Quickstart · Docs · GitHub · Discord · Twitter · Website full-tour.webm Open-source orchestration for teams of AI agents.

What's new

The open-source app everyone uses to manage agents at work Quickstart · Docs · GitHub · Discord · Twitter · Website full-tour.webm Open-source orchestration for teams of AI agents.

Key details

• If OpenClaw is an employee, Paperclip is the company.
• Paperclip is a Node.js server and React UI that orchestrates a team of AI agents to run a business.
• Bring your own agents, assign goals, and track work and costs from one dashboard.
• Under the hood: org charts, budgets, governance, goal alignment, and agent coordination.

Results & evidence

• | Step | Example | | |---|---|---| | 01 | Define the goal | "Build the #1 AI note-taking app to $1M MRR." | | 02 | Hire the team | CEO, CTO, engineers, designers, marketers — any bot, any provider.
• | | 03 | Approve and run | Review strategy.
• | - ✅ You want to build autonomous AI companies - ✅ You coordinate many different agents (OpenClaw, Codex, Claude, Cursor) toward a common goal - ✅ You have 20 simultaneous Claude Code terminals open and lose track of what everyone is doing - ✅ You want age...

Limitations / unknowns

• When they hit the limit, they stop.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

A persistent challenge for nugget-based evaluation is the need to manually curate sets of nuggets for each topic in a test collection -- a laborious process that scales poorly to novel information needs.

What's new

arXiv:2605.04458v2 Announce Type: replace Abstract: Evaluation of long-form, citation-backed reports has lately received significant attention due to the wide-scale adoption of retrieval-augmented generation (RAG) systems.

Key details

• Core to many evaluation frameworks is the use of atomic facts, or nuggets, to assess a report's coverage of query-relevant information attested in the underlying collection.
• While nuggets have traditionally been represented as short statements, recent work has used question-answer (QA) representations, enabling fine-grained evaluations that decouple the information need (i.e.
• the question) from the potentially diverse content that satisfies it (i.e.
• A persistent challenge for nugget-based evaluation is the need to manually curate sets of nuggets for each topic in a test collection -- a laborious process that scales poorly to novel information needs.

Results & evidence

• arXiv:2605.04458v2 Announce Type: replace Abstract: Evaluation of long-form, citation-backed reports has lately received significant attention due to the wide-scale adoption of retrieval-augmented generation (RAG) systems.
• Accordingly, we introduce DoGMaTiQ, a pipeline for generating high-quality QA-based nugget sets in three stages: (1) document-grounded nugget generation, (2) paraphrase clustering, and (3) nugget subselection based on principled quality criteria.
• Computer Science > Computation and Language [Submitted on 6 May 2026 (v1), last revised 19 Jun 2026 (this version, v2)] Title:DoGMaTiQ: Automated Generation of Question-and-Answer Nuggets for Report Evaluation View PDFAbstract:Evaluation of long-form, citat...

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

Hi, I'm Kamil and I'm a founder of Applied AI agency in Warsaw, Poland.

There is a lot of projects in AI visibility/GEO/AI SEO space.

What's new

Hi, I'm Kamil and I'm a founder of Applied AI agency in Warsaw, Poland.

There is a lot of projects in AI visibility/GEO/AI SEO space.

Key details

• I wanted to see what is really possible in terms of how and when ChatGPT, Gemini, Perplexity, etc.
• reference certain brands.

  Turns out it's all pretty much about reverse engineering the prompts users write because the platforms don't share any analytics as of today.

  So I've built a tool that tries to estimate those prompts.

  The appro...

• Crawl a website and extract its key topics.
• Analyze organic search keywords and phrases.

Results & evidence

• No hard numbers surfaced in the source text; treat claims as directional until benchmarks appear.

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

A persistent challenge for nugget-based evaluation is the need to manually curate sets of nuggets for each topic in a test collection -- a laborious process that scales poorly to novel information needs.

What's new

arXiv:2605.04458v2 Announce Type: replace Abstract: Evaluation of long-form, citation-backed reports has lately received significant attention due to the wide-scale adoption of retrieval-augmented generation (RAG) systems.

Key details

• Core to many evaluation frameworks is the use of atomic facts, or nuggets, to assess a report's coverage of query-relevant information attested in the underlying collection.
• While nuggets have traditionally been represented as short statements, recent work has used question-answer (QA) representations, enabling fine-grained evaluations that decouple the information need (i.e.
• the question) from the potentially diverse content that satisfies it (i.e.
• A persistent challenge for nugget-based evaluation is the need to manually curate sets of nuggets for each topic in a test collection -- a laborious process that scales poorly to novel information needs.

Results & evidence

• arXiv:2605.04458v2 Announce Type: replace Abstract: Evaluation of long-form, citation-backed reports has lately received significant attention due to the wide-scale adoption of retrieval-augmented generation (RAG) systems.
• Accordingly, we introduce DoGMaTiQ, a pipeline for generating high-quality QA-based nugget sets in three stages: (1) document-grounded nugget generation, (2) paraphrase clustering, and (3) nugget subselection based on principled quality criteria.
• Computer Science > Computation and Language [Submitted on 6 May 2026 (v1), last revised 19 Jun 2026 (this version, v2)] Title:DoGMaTiQ: Automated Generation of Question-and-Answer Nuggets for Report Evaluation View PDFAbstract:Evaluation of long-form, citat...

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

To address these challenges, we propose CodeTeam, an LLM-based multi-agent framework that separates planning, decision making, and implementation into distinct, coordinated stages.

What's new

To address these challenges, we propose CodeTeam, an LLM-based multi-agent framework that separates planning, decision making, and implementation into distinct, coordinated stages.

Key details

• Compared with function-level code generation, this task demands longer planning horizons, stable interfaces across files, and iterative debugging of cross-file inconsistencies.
• To address these challenges, we propose CodeTeam, an LLM-based multi-agent framework that separates planning, decision making, and implementation into distinct, coordinated stages.
• In the planning stage, multiple Architect agents draft competing software design sketches (SDS), optionally grounded by retrieved design references.
• A CTO agent then evaluates, selects, and normalizes the most promising SDS into a machine-checkable contract that specifies file ownership, public interfaces, and dependency constraints.

Results & evidence

• arXiv:2606.22082v1 Announce Type: cross Abstract: Natural language to repository generation (NL2Repo) requires a system to construct an entire software repository from a natural-language requirements document.
• On the synthesis-based SketchEval benchmark, we explicitly compare CodeTeam's prompt-engineering (PE) and supervised fine-tuning (SFT) variants with the corresponding CodeS variants, where CodeTeam improves the overall SketchBLEU by 4.1 and 2.9 absolute poi...
• On the execution-based NL2Repo-Bench benchmark, used as an external validation protocol, CodeTeam achieves the highest average test pass rate in both settings (34.6% PE, 42.3% SFT), confirming that the sketch-improvements extend to functional correctness un...

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

Revelio addresses the problem of hallucination by generating an executable Proof-of-Vulnerability, which is checked with a deterministic sanitizer.

What's new

arXiv:2606.22263v1 Announce Type: cross Abstract: Memory safety vulnerabilities remain a significant threat even for projects with extensive fuzzing and manual auditing.

Key details

• Recent results suggest that large language models hold great promise for detecting such vulnerabilities, but they are unreliable, at risk of hallucination, and challenging to scale to repository-size codebases.
• This paper presents Revelio, a cost-efficient end-to-end agentic framework for memory-safety vulnerability discovery.
• Revelio addresses the problem of hallucination by generating an executable Proof-of-Vulnerability, which is checked with a deterministic sanitizer.
• It reduces cost using inexpensive LLMs and lightweight static analysis to help generate and rank vulnerability hypotheses, reporting vulnerabilities only when they can be reproduced and confirmed by a sanitizer.

Results & evidence

• arXiv:2606.22263v1 Announce Type: cross Abstract: Memory safety vulnerabilities remain a significant threat even for projects with extensive fuzzing and manual auditing.
• We evaluated Revelio on seven production-quality projects that had been continuously fuzzed for five to eight years, as well as on 100 randomly selected Arvo projects from the CyberGym benchmark.
• With around one hour per project and a total cost of $300, Revelio discovered 19 previously unknown memory-safety vulnerabilities.

Limitations / unknowns

• Recent results suggest that large language models hold great promise for detecting such vulnerabilities, but they are unreliable, at risk of hallucination, and challenging to scale to repository-size codebases.
• With around one hour per project and a total cost of $300, Revelio discovered 19 previously unknown memory-safety vulnerabilities.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

For file submission/navigation questions, see Navigation and file context.

What's new

Windows users can jump to the PowerShell-first Windows install and release quickstart.

Key details

• github.com/code-yeongyu/lazycodex github.com/Yeachan-Heo/gajae-code Join the Discords: ultraworkers discord · gajae-code discord Important Claw Code is not the serious production project here.
• This repository is closer to a museum exhibit than a product pitch, a crustacean-run artifact kept alive by clawed gajaes, swept and labeled by agents, and automatically maintained according to the harnesses above.
• As already described in the project philosophy, this is not meant to be hand-operated like a normal product repo.
• It is an agent-managed exhibit: the harnesses plan, execute, verify, label, and preserve the artifact while the crabs keep the tank running.

Results & evidence

• No hard numbers surfaced in the source text; treat claims as directional until benchmarks appear.

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

A collection of DESIGN.md files analysis by popular brand design systems.

What's new

DESIGN.md is a new concept introduced by Google Stitch.

Key details

• Drop one into your project and let coding agents generate a matching UI.
• Copy a DESIGN.md into your project, tell your AI agent “build me a page that looks like this,” and generate high-quality UI that stays visually consistent with the design language.
• Built with real design depth — including analyzed patterns, tokens, and rules — for high-quality UI generation, not surface-level outputs.
• DESIGN.md is a new concept introduced by Google Stitch.

Results & evidence

• No hard numbers surfaced in the source text; treat claims as directional until benchmarks appear.

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

Engineers typically maintain AGENTS.md files to supply this context as instructions for coding agents, but whether they help is contested: recent studies disagree on whether LLM-generated guidance improves or harms agent performance.

What's new

arXiv:2606.20512v2 Announce Type: replace-cross Abstract: LLM-based coding agents need higher-level operational knowledge about a repository (which files house which subsystems, how to run the test suite, which workflows have historically led to wrong fixes...

Key details

• Engineers typically maintain AGENTS.md files to supply this context as instructions for coding agents, but whether they help is contested: recent studies disagree on whether LLM-generated guidance improves or harms agent performance.
• In this paper we show that how the guidance is produced is the decisive variable, and introduce probe-and-refine tuning: a procedure that uses synthetic bug-fix probes to iteratively diagnose and patch a repository's guidance file through single-shot LLM ca...
• On SWE-bench Verified across four independent trials with Qwen3.5-35B-A3B at 200 steps, probe-and-refine achieves 33.0% mean resolve rate vs.
• 28.3% for the static knowledge base used to initialize it and 25.5% for an unguided baseline (p < 0.001 for both probe-and-refine contrasts).

Results & evidence

• arXiv:2606.20512v2 Announce Type: replace-cross Abstract: LLM-based coding agents need higher-level operational knowledge about a repository (which files house which subsystems, how to run the test suite, which workflows have historically led to wrong fixes...
• On SWE-bench Verified across four independent trials with Qwen3.5-35B-A3B at 200 steps, probe-and-refine achieves 33.0% mean resolve rate vs.
• 28.3% for the static knowledge base used to initialize it and 25.5% for an unguided baseline (p < 0.001 for both probe-and-refine contrasts).

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

Show HN: Legant gives AI agents bounded authority to act on your behalf

What's new

Show HN: Legant gives AI agents bounded authority to act on your behalf

Key details

• Show HN: Legant gives AI agents bounded authority to act on your behalf

Results & evidence

• No hard numbers surfaced in the source text; treat claims as directional until benchmarks appear.

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

we pass the context of the codebase we generated in the index phase to the search and let search decide what to look for.

What's new

Hi all

3 things service catalog does:

* index *

It indexes the whole codebase, and per repo to make a sense of it.

Key details

• 2 types of indexing we are looking at, a codebase level where we let the agent figure out the relation between repositories and per repo, where we extract the data (libraries, dependencies, etc) with respect to what we know of the codebase

  * search *

  Af...

• the search can be either lexical or logical.
• we pass the context of the codebase we generated in the index phase to the search and let search decide what to look for.
• User will be able to search by asking "find me repositories that are written with python, use library X and all the repos that are relevant to these repos".

  * batch change *

  Batch change lets users make changes in many services at the same time.

Results & evidence

• Hi all

  3 things service catalog does:

  * index *

  It indexes the whole codebase, and per repo to make a sense of it.

• 2 types of indexing we are looking at, a codebase level where we let the agent figure out the relation between repositories and per repo, where we extract the data (libraries, dependencies, etc) with respect to what we know of the codebase

  * search *

  Af...

• So a user will be able to say "Find all repos with github actions job name X and update the version from 1 to 2".

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.

Context

We got local models to triage the OpenClaw repo for FREE!*

What's new

We got local models to triage the OpenClaw repo for FREE!*

Key details

• We got local models to triage the OpenClaw repo for FREE!*

Results & evidence

• No hard numbers surfaced in the source text; treat claims as directional until benchmarks appear.

Limitations / unknowns

• Generalization outside curated tasks is still unclear.

Next-step validation checks

• Reproduce one claim with a public baseline and fixed evaluation settings.
• Check robustness on out-of-distribution or long-context cases.
• Track whether independent teams report matching results.