Radar

Run status: ok Rule: read-only. No login, post, DM, outreach, deploy, credentials, model-route changes, or DNS. Firecrawl only under explicit bounded approved budget. YouTube: ok | source_group=ai_youtube_data_api_v3 | requests=1 | results=10 Firecrawl: skipped_by_owner_decision | priority=4 residual-only | requests_used=0/25 | credits_used=0 | monthly_budget_usd=25 Sources scanned: 16 | fetches_ok=32 | fetches_error=5 Model layers: Gemma=40 | Qwen tools-only=35 | Terra final=35 | errors=0 Candidates: AI=5 business=0 | new_this_cycle=5 | escalated=5 - [arxiv] score=9.0 installable=False — Quantifying Overclaiming Propensity in Frontier LLM Agents - URL: http://arxiv.org/abs/2609.20812v1 - Why: New benchmark exposes agent completion overclaims, directly relevant to TNMC autonomous-agent trust controls. - Gemma: {'status': 'ok', 'verdict': 'KEEP', 'score': 9, 'reason': 'Specific, novel metric quantifying LLM overclaiming propensity with concrete evidence.', 'model': '/Users/macstudio4/models/gemma-4-e4b-gguf/gemma-4-E4B-it-Q4_K_M.gguf', 'latency_ms': 885.3, 'content_sha256': 'cb2da38d58d1eaed2df2d591f2bd0cb5336448486346cc76f8d179a2ab092025', 'attempts': [{'attempt': 1, 'status': 'ok', 'latency_ms': 885.3}]} - Qwen checks: {'status': 'ok', 'verdict': 'KEEP', 'score': 9, 'checks': {'freshness_ok': True, 'source_allowed': True, 'spam_like': False, 'evidence_level': 3}, 'health': {'ok': True, 'status': 200, 'latency_ms': 3.0, 'payload': {'status': 'ok', 'model': '/Users/studioforge/models/Qwen3-Coder-30B-A3B-Instruct-4bit-official', 'policy': 'tools_only'}}, 'model': '/Users/studioforge/models/Qwen3-Coder-30B-A3B-Instruct-4bit-official', 'latency_ms': 820.7, 'attempts': 1, 'tool_call_id': 'call_e37972aba91947d9b28f2489'} - Terra: {'status': 'ok', 'verdict': 'KEEP', 'score': 9, 'summary': 'New benchmark exposes agent completion overclaims, directly relevant to TNMC autonomous-agent trust controls.', 'reason': 'Specific frontier-agent reliability research addresses a critical operational risk, though it lacks immediate buyer intent.', 'model': 'gpt-5.6-terra', 'provider': 'openai-codex', 'actor': 'world-scanner-terra-final', 'latency_ms': 13538.2, 'usage_receipt': '/Users/macstudio4/vault-studio2/operations/plan-maestro-cierre-20260722/M-B-2-20260918T081825Z/terra-ad11e9ef5b60f8ac.usage.json', 'usage_receipt_sha256': 'aa848303fedfe32a93c6c6de4d55323bc71f9411dcea969e7de98ab303136f42', 'content_sha256': '60d5bb00d075639c50eba4caad0032e6d70d104d6b2727850692c43b7eda2470'} - [hackernews] score=8.2 installable=False — Show HN: MCPJam - the first testing & evaluations platform for MCP servers - URL: https://www.mcpjam.com - Why: MCPJam targets enterprise testing and observability for MCP server integrations across major AI clients. - Gemma: {'status': 'ok', 'verdict': 'KEEP', 'score': 8, 'reason': 'Specific platform for testing AI server interactions with concrete URL.', 'model': '/Users/macstudio4/models/gemma-4-e4b-gguf/gemma-4-E4B-it-Q4_K_M.gguf', 'latency_ms': 946.4, 'content_sha256': '09ee08dd6a6767983393d994b57bdede9b0d950107085056704d5de654727180', 'attempts': [{'attempt': 1, 'status': 'ok', 'latency_ms': 946.4}]} - Qwen checks: {'status': 'ok', 'verdict': 'KEEP', 'score': 9, 'checks': {'freshness_ok': True, 'source_allowed': True, 'spam_like': False, 'evidence_level': 3}, 'health': {'ok': True, 'status': 200, 'latency_ms': 3.1, 'payload': {'status': 'ok', 'model': '/Users/studioforge/models/Qwen3-Coder-30B-A3B-Instruct-4bit-official', 'policy': 'tools_only'}}, 'model': '/Users/studioforge/models/Qwen3-Coder-30B-A3B-Instruct-4bit-official', 'latency_ms': 812.9, 'attempts': 1, 'tool_call_id': 'call_ea203ec2455e4f17b553a9d7'} - Terra: {'status': 'ok', 'verdict': 'KEEP', 'score': 8, 'summary': 'MCPJam targets enterprise testing and observability for MCP server integrations across major AI clients.', 'reason': 'Specific emerging infrastructure need with clear enterprise pain, credible operator background, and practical relevance to TNMC AI integration work.', 'model': 'gpt-5.6-terra', 'provider': 'openai-codex', 'actor': 'world-scanner-terra-final', 'latency_ms': 12096.1, 'usage_receipt': '/Users/macstudio4/vault-studio2/operations/plan-maestro-cierre-20260722/M-B-2-20260918T081825Z/terra-b0e77715be37c21c.usage.json', 'usage_receipt_sha256': 'b52de837a019857905c2304354e8b6269e5d688a2608220a09376aedafcec977', 'content_sha256': 'a2ef2a6ef1266805e84563abbe8d2c7ffafd1583b6fc31a0f4ef5ee9dd051d98'} - [hackernews] score=8.2 installable=False — Show HN: Craigslist for agent skills, curated by a human - URL: https://skillbay.sh/ - Why: Human-curated marketplace tests paid AI-agent skills for practical business workflows. - Gemma: {'status': 'ok', 'verdict': 'KEEP', 'score': 8, 'reason': 'Specific new platform for buying AI agent skills with concrete evidence.', 'model': '/Users/macstudio4/models/gemma-4-e4b-gguf/gemma-4-E4B-it-Q4_K_M.gguf', 'latency_ms': 845.5, 'content_sha256': 'f956a042e6315c3fb877f9a5ddf6535f3cfcdc175dbe702d9c1f8843b4e997e0', 'attempts': [{'attempt': 1, 'status': 'ok', 'latency_ms': 845.5}]} - Qwen checks: {'status': 'ok', 'verdict': 'KEEP', 'score': 9, 'checks': {'freshness_ok': True, 'source_allowed': True, 'spam_like': False, 'evidence_level': 3}, 'health': {'ok': True, 'status': 200, 'latency_ms': 4.9, 'payload': {'status': 'ok', 'model': '/Users/studioforge/models/Qwen3-Coder-30B-A3B-Instruct-4bit-official', 'policy': 'tools_only'}}, 'model': '/Users/studioforge/models/Qwen3-Coder-30B-A3B-Instruct-4bit-official', 'latency_ms': 826.4, 'attempts': 1, 'tool_call_id': 'call_a03c7bc8afb94ac4a04a88f3'} - Terra: {'status': 'ok', 'verdict': 'KEEP', 'score': 7, 'summary': 'Human-curated marketplace tests paid AI-agent skills for practical business workflows.', 'reason': 'Concrete new monetization platform; relevant to TNMC agent services, though demand and traction remain unverified.', 'model': 'gpt-5.6-terra', 'provider': 'openai-codex', 'actor': 'world-scanner-terra-final', 'latency_ms': 17302.5, 'usage_receipt': '/Users/macstudio4/vault-studio2/operations/plan-maestro-cierre-20260722/M-B-2-20260918T081825Z/terra-a3dd8386ec71dbb0.usage.json', 'usage_receipt_sha256': '20f45f9f27c58f050e9a7ebd6b7719c07f3256a7291c87d4742c707ae56f0160', 'content_sha256': '9b8549920505dcb57fa0007e1cead6f5a378d9df62939f15556f513321485b1c'} - [hackernews] score=7.9 installable=False — Show HN: Multiplayer Mode for AI Agents - URL: https://gotincan.com/ - Why: New encrypted real-time coordination layer for multi-harness AI agent teams. - Gemma: {'status': 'ok', 'verdict': 'KEEP', 'score': 8, 'reason': 'Specific, functional AI agent coordination tool with public demo.', 'model': '/Users/macstudio4/models/gemma-4-e4b-gguf/gemma-4-E4B-it-Q4_K_M.gguf', 'latency_ms': 855.1, 'content_sha256': '53aedb15261d3b8e4aca6d76471f438ab4687e1bae2cd98f78a703c406bb49bf', 'attempts': [{'attempt': 1, 'status': 'ok', 'latency_ms': 855.1}]} - Qwen checks: {'status': 'ok', 'verdict': 'KEEP', 'score': 9, 'checks': {'freshness_ok': True, 'source_allowed': True, 'spam_like': False, 'evidence_level': 3}, 'health': {'ok': True, 'status': 200, 'latency_ms': 4.6, 'payload': {'status': 'ok', 'model': '/Users/studioforge/models/Qwen3-Coder-30B-A3B-Instruct-4bit-official', 'policy': 'tools_only'}}, 'model': '/Users/studioforge/models/Qwen3-Coder-30B-A3B-Instruct-4bit-official', 'latency_ms': 771.0, 'attempts': 1, 'tool_call_id': 'call_8db88a68bf694555b83ab25a'} - Terra: {'status': 'ok', 'verdict': 'KEEP', 'score': 8, 'summary': 'New encrypted real-time coordination layer for multi-harness AI agent teams.', 'reason': 'Specific public product addresses TNMC’s multi-agent orchestration needs, though traction and operational reliability remain unverified.', 'model': 'gpt-5.6-terra', 'provider': 'openai-codex', 'actor': 'world-scanner-terra-final', 'latency_ms': 10496.5, 'usage_receipt': '/Users/macstudio4/vault-studio2/operations/plan-maestro-cierre-20260722/M-B-2-20260918T081825Z/terra-9881067f0232e208.usage.json', 'usage_receipt_sha256': 'cceb841c4456de76427add610c5adde4e01fca00a09eb6546648c56a580a1b6e', 'content_sha256': '6b4ce537a8bf4ad0843d53e985ed33efe1643c076b21a1eab0ed09a141758b6f'} - [hackernews] score=7.9 installable=False — Show HN: AutoBot – live voice control for long-running AI work - URL: https://github.com/demeyer1/Autobot - Why: Open-source voice-controlled AI harness posts benchmarked long-running workflow performance and privacy-focused local persistence. - Gemma: {'status': 'ok', 'verdict': 'KEEP', 'score': 8, 'reason': 'Specific project demonstrating voice control for long-running AI tasks with benchmarks.', 'model': '/Users/macstudio4/models/gemma-4-e4b-gguf/gemma-4-E4B-it-Q4_K_M.gguf', 'latency_ms': 960.8, 'content_sha256': '547282e3eb66ae139668bccddaa2f1ea85bd9d01177437b749a6296f596fc374', 'attempts': [{'attempt': 1, 'status': 'ok', 'latency_ms': 960.8}]} - Qwen checks: {'status': 'ok', 'verdict': 'KEEP', 'score': 9, 'checks': {'freshness_ok': True, 'source_allowed': True, 'spam_like': False, 'evidence_level': 3}, 'health': {'ok': True, 'status': 200, 'latency_ms': 4.5, 'payload': {'status': 'ok', 'model': '/Users/studioforge/models/Qwen3-Coder-30B-A3B-Instruct-4bit-official', 'policy': 'tools_only'}}, 'model': '/Users/studioforge/models/Qwen3-Coder-30B-A3B-Instruct-4bit-official', 'latency_ms': 786.5, 'attempts': 1, 'tool_call_id': 'call_a8676a3917204e809c495b28'} - Terra: {'status': 'ok', 'verdict': 'KEEP', 'score': 8, 'summary': 'Open-source voice-controlled AI harness posts benchmarked long-running workflow performance and privacy-focused local persistence.', 'reason': 'Verified repository supplies concrete benchmarks and operational patterns relevant to TNMC’s autonomous-agent reliability and voice-interface evaluation.', 'model': 'gpt-5.6-terra', 'provider': 'openai-codex', 'actor': 'world-scanner-terra-final', 'latency_ms': 42315.7, 'usage_receipt': '/Users/macstudio4/vault-studio2/operations/plan-maestro-cierre-20260722/M-B-2-20260918T081825Z/terra-a24a79122f64dc25.usage.json', 'usage_receipt_sha256': '740427aa3a67bf12b9772b950eb17e33f9884ede2f3f34d15580655a70d6616e', 'content_sha256': '36c0d8c02a5d9810d690401ee3a8ab0e04b759d29e867ad9b9bfc9ca3839ea71'} - None above threshold this cycle. - ok creative_marketing_trade_press 200 181.1ms https://www.marketingdive.com/feeds/news/ - ok creative_marketing_trade_press 200 222.2ms https://www.adweek.com/feed/ - ok creative_marketing_trade_press 200 126.2ms https://digiday.com/feed/ - ok creative_agencies_and_studios 200 474.9ms https://www.creativeboom.com/feed/ - ok creative_platform_product_launches 200 275.9ms https://blog.google/products/ads-commerce/rss/ - ok creative_platform_product_launches 200 247.5ms https://blog.youtube/rss/ - ok ai_hackernews_algolia 200 714.2ms https://hn.algolia.com/api/v1/search_by_date?query=Show+HN+AI+agent&tags=story&hitsPerPage=12 - ok ai_hackernews_algolia 200 534.6ms https://hn.algolia.com/api/v1/search_by_date?query=open+source+LLM+tool&tags=story&hitsPerPage=12 - ok ai_hackernews_algolia 200 487.7ms https://hn.algolia.com/api/v1/search_by_date?query=local+LLM+MLX+Qwen+GLM&tags=story&hitsPerPage=12 - ok ai_github_trending 200 2036.3ms https://github.com/trending?spoken_language_code=en - ok ai_arxiv_cs 200 243.1ms https://export.arxiv.org/api/query?search_query=cat%3Acs.CL+OR+cat%3Acs.AI&sortBy=submittedDate&sortOrder=descending&max_results=12 - ok ai_github_org_api 200 11408.0ms https://api.github.com/orgs/openai/repos?type=public&sort=updated&direction=desc&per_page=10 - ok ai_github_org_api 200 2274.7ms https://api.github.com/orgs/anthropics/repos?type=public&sort=updated&direction=desc&per_page=10 - ok ai_github_org_api 200 1048.9ms https://api.github.com/orgs/google-deepmind/repos?type=public&sort=updated&direction=desc&per_page=10 - ok ai_github_org_api 200 730.8ms https://api.github.com/orgs/meta-llama/repos?type=public&sort=updated&direction=desc&per_page=10 - ok ai_github_org_api 200 678.4ms https://api.github.com/orgs/mistralai/repos?type=public&sort=updated&direction=desc&per_page=10 - ok ai_github_org_api 200 619.7ms https://api.github.com/orgs/QwenLM/repos?type=public&sort=updated&direction=desc&per_page=10 - ok ai_github_org_api 200 585.3ms https://api.github.com/orgs/zai-org/repos?type=public&sort=updated&direction=desc&per_page=10 - ok ai_github_org_api 200 617.7ms https://api.github.com/orgs/ml-explore/repos?type=public&sort=updated&direction=desc&per_page=10 - ok ai_github_org_api 200 629.7ms https://api.github.com/orgs/deepseek-ai/repos?type=public&sort=updated&direction=desc&per_page=10 - ok ai_huggingface_trending_models 200 158.6ms https://huggingface.co/api/models?sort=trendingScore&direction=-1&limit=30 - ok ai_huggingface_trending_papers 200 275.5ms https://huggingface.co/api/daily_papers?sort=trending&limit=30 - ok ai_github_permanent_release_watch 200 720.9ms https://api.github.com/repos/block/buzz/releases?per_page=10 - ok ai_youtube_data_api_v3 200 412.5ms https://www.googleapis.com/youtube/v3/search - ok business_hackernews_algolia 200 600.4ms https://hn.algolia.com/api/v1/search_by_date?query=need+landing+page&tags=story%2Ccomment&hitsPerPage=15 - ok business_hackernews_algolia 200 1107.3ms https://hn.algolia.com/api/v1/search_by_date?query=looking+for+designer&tags=story%2Ccomment&hitsPerPage=15 - ok business_hackernews_algolia 200 373.4ms https://hn.algolia.com/api/v1/search_by_date?query=brand+identity&tags=story%2Ccomment&hitsPerPage=15 - ok business_hackernews_algolia 200 351.7ms https://hn.algolia.com/api/v1/search_by_date?query=local+SEO+small+business&tags=story%2Ccomment&hitsPerPage=15 - ok business_reddit_rss 200 536.2ms https://www.reddit.com/r/forhire/new.rss?limit=25 - error business_reddit_rss 429 266.9ms https://www.reddit.com/r/smallbusiness/new.rss?limit=25

This note describes the Scout run. Candidate titles and URLs come only from the World Scanner JSON above.