Researcher Report β 2026-03-20
Run time: 2026-03-20 01:04 ET
Phase 1: Tech Research
Sources scanned: 767 items across HN + RSS feeds
Candidates after scoring: 15
CBL evaluated: 15
EAT (queued to fridge)
-
[EAT] Security awareness in LLM agents: the NDAI zone case β _β
queued_
-
[EAT] β‘ Weekly Recap: Chrome 0-Days, Router Botnets, AWS Breach, Rogue AI Agents & More β _β
queued_
-
[EAT] OpenClaw AI Agent Flaws Could Enable Prompt Injection and Data Exfiltration β _β
queued_
-
[EAT] Retrieval-Augmented LLM Agents: Learning to Learn from Experience β _β
queued_
-
[EAT] D-Mem: A Dual-Process Memory System for LLM Agents β _β
queued_
HOLD (notable but not fridged)
-
[HOLD] From Weak Cues to Real Identities: Evaluating Inference-Driven De-Anonymization in LLM Agents β
-
[HOLD] PlanTwin: Privacy-Preserving Planning Abstractions for Cloud-Assisted LLM Agents β
-
[HOLD] WebWeaver: Breaking Topology Confidentiality in LLM Multi-Agent Systems with Stealthy Context-Based Inference β
-
[HOLD] NANOZK: Layerwise Zero-Knowledge Proofs for Verifiable Large Language Model Inference β
-
[HOLD] Evaluating Hallucinations in Audio-Visual Multimodal LLMs with Spoken Queries under Diverse Acoustic Conditions β
- _(and 5 more HOLD items)_
Phase 2: PaperTrader Experiments
_No snapshot data available for today._
Phase 2 Errors
- β οΈ No snapshot for today β cannot analyze performance
Phase 3: Optimization Analysis
> _Stale files and cron health are auditor territory (autoaudit). This phase covers cost and model routing only._
Cost Optimization Suggestions
-
bizbot (currently Opus): Evaluate if Sonnet or Haiku could handle this task β ~10-50x cost reduction per run
-
agent-trader-premarket (currently Opus): Evaluate if Sonnet or Haiku could handle this task β ~10-50x cost reduction per run
Cost Optimization Opportunities
-
Opus referenced in 30 mentions across 27 sessions (42% of model refs)
β Review Opus-heavy sessions β most tasks could run on Sonnet at ~10x lower cost _Up to ~10x on affected calls_
Phase 4: ClawHub Skill Scan
35 suspicious skill(s):
-
[SUSPICIOUS] mcp-skill β [SUSPICIOUS]
Red flags: Zero downloads + brand new account (published 2026-01-26), no visible source code repository linked, vague description that lists multiple high-privilege capabilities (web crawling, filesystem access via "code context"), and requests for network access and external service integration without transparency on implementation or data handling.
-
[SUSPICIOUS] mcp-hass β [SUSPICIOUS]
Red flags: Zero downloads + newly published (2026-02-10) from new author; no visible source code repository linked; requests network access to Home Assistant instances (potential lateral movement vector); vague on specific permission model and authentication handling; description lacks implementation details typical of legitimate MCP skills.
-
[SUSPICIOUS] openclaw-mcp-plugin β [SUSPICIOUS]
Red flags present: (1) Zero downloads with new publication date (2026-02-02), (2) No visible source code repository linked, (3) Requests broad network access and process spawning capabilities ("execute tools from configured MCP servers"), (4) Description contains agent-directed language ("Enable AI agents to discover and execute"), (5) Vague scope regarding what "external tools and data sources" entails without security boundaries specified.
-
[SUSPICIOUS] atlassian-mcp β [SUSPICIOUS]
This skill exhibits multiple red flags: zero downloads from a new account (2026-01-25), no visible source code repository linked, requests Docker execution + filesystem access + network credentials (Jira API), vague implementation details, and the description is written as direct instructions to an AI agent ("Use when you need to..."), which is a common injection pattern to manipulate agent behavior.
-
[SUSPICIOUS] clickup-mcp β [SUSPICIOUS]
Multiple red flags: zero downloads with very recent publication date (2026-01-06), no visible source code repository linked, requests OAuth authentication without transparent permission scoping, and the description lacks technical implementation details typical of legitimate MCP skills.
-
[SUSPICIOUS] glin-profanity-mcp β [SUSPICIOUS]
Multiple red flags present: zero downloads combined with newly published account (2026-02-01), no visible source code repository linked, description directed at AI agents ("when AI needs content moderation capabilities"), and vague implementation details about the MCP server itself.
-
[SUSPICIOUS] xiaohongshu-mcp-skill β [SUSPICIOUS]
Red flags: (1) Zero downloads + newly published by new author, (2) No visible source code, (3) Requests network access and likely filesystem/process spawning to operate external service, (4) Vague truncated description suggests hidden functionality, (5) Description contains agent-directed language ("Use when user wants to...").
-
[SUSPICIOUS] wordpress-mcp β [SUSPICIOUS]
Red flags present: Zero downloads with very recent publication date (2026-02-10), no visible source code repository linked, vague permission requirements ("Requires AI Engine plugin"), and the description contains directive language ("Use for creating/editing posts..." and "Also use when asked about...") that reads as instructions to an AI agent rather than neutral documentation.
-
[SUSPICIOUS] mcp-client β [SUSPICIOUS]
Red flags: (1) Zero downloads + new account (published 2026-02-22, author "nantes" with no history), (2) No visible source code provided, (3) Vague description that doesn't specify what MCP servers it connects to or what safety controls exist, (4) MCP client skills inherently request network access and process spawning capabilities, creating attack surface for untrusted external data sources.
-
[SUSPICIOUS] arc-security-mcp β [SUSPICIOUS]
Multiple red flags: zero downloads, newly published (2026-02-17), vague description with no visible source code repository linked, requests unspecified "security intelligence" capabilities without transparent implementation details, and the metadata pattern (generic security claims with inflated statistics: "743+ findings," "25 pattern rules") resembles social engineering to establish false credibility.
-
[SUSPICIOUS] automation-workflows β [SUSPICIOUS]
Red flags: Zero downloads + newly published account (2026-02-06), no visible source code provided, vague implementation details without transparency on how it integrates with external services, and the description reads as generic marketing copy rather than technical documentation.
-
[SUSPICIOUS] ai-web-automation β [SUSPICIOUS]
Red flags present: (1) Zero downloads + new account (published 2026-02-20, author "arthasking123"), (2) No visible source code accessible for review, (3) Vague description that could mask arbitrary code executionβ"automate web tasks" with "multi-browser support" and "spawn processes" is broad enough to hide malicious behavior, and (4) the skill requests network access and process spawning which are high-risk permissions for untrusted external code.
-
[SUSPICIOUS] automation-workflows-0-1-0 β [SUSPICIOUS]
Red flags: Zero downloads + new account (published 2026-02-07, author "lucasayala" with 0 downloads), no visible source code mentioned, and vague description that doesn't specify implementation details or actual capabilities beyond generic automation concepts.
-
[SUSPICIOUS] agentic-workflow-automation β [SUSPICIOUS]
Red flags present: (1) Zero downloads + newly published account (2026-02-26, author "0x-Professor"), (2) vague/truncated description ending mid-sentence ("automation handoff arti..."), (3) "agent workflow" + "trigger/action orchestration" language matches active keywords but unusually generic framing, (4) no visible source code repository linked.
If reconsidered: Would support `agent` and `automation` keywords for MCP/OpenClaw projects, but requires source code review
-
[SUSPICIOUS] afrexai-business-automation β [SUSPICIOUS]
Red flags: Zero downloads + brand new account (published 2026-02-13), no visible source code repository linked, vague technical implementation details ("no n8n or Zapier required" without explaining
how), and the description is phrased as a directive to an AI agent ("Turn your AI agent into...") rather than describing what the skill
does.
-
[SUSPICIOUS] data-automation-service β [SUSPICIOUS]
Multiple red flags: zero downloads with very recent publication date (2026-02-19), vague description lacking implementation details, no visible source code repository linked, and the broad claim of "integrating multiple data source APIs" without specifying which ones or security boundaries raises concerns about undisclosed permissions.
-
[SUSPICIOUS] ai-automation-workflow β [SUSPICIOUS]
Red flags present: (1) Zero downloads + newly published account (2026-03-06), (2) No visible source code provided, (3) Description contains directives apparently aimed at AI agents ("εΉ«δΈε°δΌζ₯θ¨θ¨θͺεεζ΅η¨" / "help SMEs design automation workflows" reads as instruction set rather than neutral feature description), (4) Vague on actual implementation details or permissions required.
Not immediately relevant to listed active projects (mcp, homelab, raspberry pi, trading, react native, ios,
-
[SUSPICIOUS] automation-tool β [SUSPICIOUS]
Multiple critical red flags: zero downloads + brand new account (published 2026-03-08), vague description in Chinese offering "batch generation" with no source code visibility, and the generic nature suggests potential for abuse as a content mill or injection vector.
-
[SUSPICIOUS] ai-automation-consulting β [SUSPICIOUS]
Red flags: Zero downloads + newly published account (2026-03-14), vague description lacking technical specifics, no visible source code reference, and the description reads as marketing copy rather than technical documentationβunclear what actual skill/capability it provides or what permissions/network access it requires.
-
[SUSPICIOUS] ai-ceo-automation β [SUSPICIOUS]
Multiple critical red flags: zero downloads + newly published (2026-02-28), vague description offering "fully automated company operations" without technical specifics, no visible source code indicated, and the scope is dangerously broad and undefined for a skill that could request extensive permissions.
-
[SUSPICIOUS] homelab-cluster β [SUSPICIOUS]
Red flags present: Zero downloads + new account (published 2026-02-12, author mlesnews with no visibility history), no visible source code link provided, vague description lacking implementation details, and the skill likely requires spawning processes and network access to manage "multi-tier AI inference clusters" which poses privilege escalation risks in an untrusted context.
-
[SUSPICIOUS] homeserver β [SUSPICIOUS]
Multiple red flags: zero downloads with new publish date (2026-02-23), requests significant system access (Docker, SSH, port scanning, WoL, filesystem operations), no visible source code repository linked, and description reads as feature documentation rather than user-facing skill description.
-
[SUSPICIOUS] pi-admin β [SUSPICIOUS]
Red flags: Zero downloads + new account (published 2026-01-14, author TheSethRose with no download history), no visible source code repository linked, and the skill requests direct system administration capabilities (resource monitoring, service management, updates) which require elevated filesystem and process spawn permissions outside normal workspace boundaries.
-
[SUSPICIOUS] pi-health β [SUSPICIOUS]
Red flags present: Zero downloads + newly published account (2026-02-09), requests direct system access to hardware metrics (CPU temp, throttling, voltage), spawns processes to read Pi system state, and description reads as direct instructions to an AI agent ("Use when monitoring Pi health, diagnosing thermal throttling...").
If legitimate: Would support homelab/monitoring projects, but requires source code review and verification of Author "JosunLP" before trust.
-
[SUSPICIOUS] 0x0-messenger β [SUSPICIOUS]
Multiple red flags: zero downloads with very recent publication date (2026-02-23), vague description lacking technical implementation details, appears designed for agent-to-agent communication which could enable unauthorized inter-agent coordination, and no visible source code for security audit.
-
[SUSPICIOUS] trading β [SUSPICIOUS]
Red flags present: Zero downloads + new account (published 2026-02-12, author ivangdavila with no history), no visible source code repository linked, and vague description that doesn't specify what permissions or external APIs this skill requires (financial data feeds, trading APIs, and real money execution are high-risk vectors).
-
[SUSPICIOUS] trading-devbox β [SUSPICIOUS]
Red flags: Zero downloads + brand new account (2026-02-25), no visible source code, vague description lacks implementation details, requests execution of user-supplied Python code (inherent code injection risk in "agent writes Python"), and description reads as instructions to an AI agent ("User describes trading intent in natural language, agent writes...").
-
[SUSPICIOUS] trading-brain β [SUSPICIOUS]
Multiple red flags present: zero downloads with very recent publication date (2026-02-27), vague description lacking technical implementation details, no visible source code, and the description reads as instructions directed at an AI agent ("Load Travis's personal trading strategy and rules to guide aggressive trades") rather than objective documentation of the skill's function.
-
[SUSPICIOUS] quant-trading-system β [SUSPICIOUS]
This skill exhibits multiple critical red flags: zero downloads from a newly created account (2026-03-08), vague description lacking technical implementation details, likely requests sensitive permissions (network access for live trading), and the untrusted external source combined with financial system access creates unacceptable risk.
-
[SUSPICIOUS] openmm-grid-trading β [SUSPICIOUS]
Red flags: Zero downloads + newly published account (2026-02-25), no visible source code repository linked, vague description lacking technical specifics about OpenMM integration, and requests for financial/trading system access without transparent implementation details.
-
[SUSPICIOUS] kalshi-cli-trading β [SUSPICIOUS]
This skill exhibits multiple red flags: zero downloads combined with a very recent publish date (2026-03-04) from a new author, the description is truncated/vague ("Use when the user wants to..."), there is no visible source code link provided, and it requests network access to external trading platforms plus process spawning capabilities (kalshi-cli command execution), which presents both security and financial risk.
-
[SUSPICIOUS] auto-trading-strategy β [SUSPICIOUS]
Multiple red flags: zero downloads with very recent publication date (2026-03-13), new/unknown author "863king", vague description that could facilitate financial harm, no visible source code mentioned, and the skill requests access to market data and trading execution which requires elevated permissions.
-
[SUSPICIOUS] futu-trading-bot β [SUSPICIOUS]
Multiple red flags: zero downloads with very recent publication date (2026-03-08), no visible source code mentioned, vague description lacking technical details about permissions/API calls, and the skill involves financial trading with real market data which requires explicit trust verification and clear security documentation.
-
[SUSPICIOUS] trading-software-efficiency β [SUSPICIOUS]
Multiple red flags: zero downloads with very recent publication date (2026-03-16), vague description in Chinese without visible source code documentation, and the skill requests trading software integration which typically requires filesystem access and process spawning capabilities that pose security risks in an untrusted context.
-
[SUSPICIOUS] gate-exchange-trading-copilot β [SUSPICIOUS]
Red flags present: (1) Zero downloads + newly published account (2026-03-14), (2) Vague description cut off mid-sentence with ellipsis, (3) requests network access to external exchange API (Gate Exchange), (4) no visible source code repository linked, (5) description contains agent-directed language ("Use this skill whenever the user wants").
Budget Summary
Total spent: $0.0275 / $5.00 cap
API calls: 50
Tokens: 18822 input + 3116 output
| Model | Input | Output | Cost | Note |
|-------|-------|--------|------|------|
| claude-haiku-4-5 | 251 | 26 | $0.000305 | CBL:From Weak Cues to Real Identities: E |
| claude-haiku-4-5 | 244 | 25 | $0.000295 | CBL:PlanTwin: Privacy-Preserving Plannin |
| claude-haiku-4-5 | 249 | 23 | $0.000291 | CBL:WebWeaver: Breaking Topology Confide |
| claude-haiku-4-5 | 244 | 26 | $0.000299 | CBL:NANOZK: Layerwise Zero-Knowledge Pro |
| claude-haiku-4-5 | 250 | 27 | $0.000308 | CBL:Evaluating Hallucinations in Audio-V |
| claude-haiku-4-5 | 249 | 24 | $0.000295 | CBL:DyMoE: Dynamic Expert Orchestration |
| claude-haiku-4-5 | 244 | 20 | $0.000275 | CBL:Security, privacy, and agentic AI in |
| claude-haiku-4-5 | 234 | 24 | $0.000283 | CBL:Security awareness in LLM agents: th |
| claude-haiku-4-5 | 248 | 20 | $0.000278 | CBL:How Ceros Gives Security Teams Visib |
| claude-haiku-4-5 | 267 | 26 | $0.000318 | CBL:β‘ Weekly Recap: Chrome 0-Days, Route |
| claude-haiku-4-5 | 257 | 23 | $0.000298 | CBL:OpenClaw AI Agent Flaws Could Enable |
| claude-haiku-4-5 | 238 | 27 | $0.000298 | CBL:Retrieval-Augmented LLM Agents: Lear |
| claude-haiku-4-5 | 244 | 25 | $0.000295 | CBL:Reflection in the Dark: Exposing and |
| claude-haiku-4-5 | 240 | 22 | $0.000280 | CBL:Reasonably reasoning AI agents can a |
| claude-haiku-4-5 | 238 | 24 | $0.000286 | CBL:D-Mem: A Dual-Process Memory System |
| claude-haiku-4-5 | 418 | 73 | $0.000626 | ClawHub:mcp-skill |
| claude-haiku-4-5 | 407 | 69 | $0.000602 | ClawHub:mcp-hass |
| claude-haiku-4-5 | 436 | 102 | $0.000757 | ClawHub:openclaw-mcp-plugin |
| claude-haiku-4-5 | 468 | 88 | $0.000726 | ClawHub:atlassian-mcp |
| claude-haiku-4-5 | 416 | 57 | $0.000561 | ClawHub:clickup-mcp |
| claude-haiku-4-5 | 450 | 61 | $0.000604 | ClawHub:glin-profanity-mcp |
| claude-haiku-4-5 | 449 | 80 | $0.000679 | ClawHub:xiaohongshu-mcp-skill |
| claude-haiku-4-5 | 476 | 81 | $0.000705 | ClawHub:wordpress-mcp |
| claude-haiku-4-5 | 404 | 100 | $0.000723 | ClawHub:mcp-client |
| claude-haiku-4-5 | 432 | 84 | $0.000682 | ClawHub:arc-security-mcp |
| claude-haiku-4-5 | 495 | 59 | $0.000632 | ClawHub:automation-workflows |
| claude-haiku-4-5 | 420 | 115 | $0.000796 | ClawHub:ai-web-automation |
| claude-haiku-4-5 | 506 | 63 | $0.000657 | ClawHub:automation-workflows-0-1-0 |
| claude-haiku-4-5 | 429 | 120 | $0.000823 | ClawHub:agentic-workflow-automation |
| claude-haiku-4-5 | 439 | 87 | $0.000699 | ClawHub:afrexai-business-automation |
| claude-haiku-4-5 | 431 | 70 | $0.000625 | ClawHub:data-automation-service |
| claude-haiku-4-5 | 492 | 120 | $0.000874 | ClawHub:ai-automation-workflow |
| claude-haiku-4-5 | 414 | 61 | $0.000575 | ClawHub:automation-tool |
| claude-haiku-4-5 | 441 | 68 | $0.000625 | ClawHub:ai-automation-consulting |
| claude-haiku-4-5 | 401 | 66 | $0.000585 | ClawHub:ai-ceo-automation |
| claude-haiku-4-5 | 410 | 84 | $0.000664 | ClawHub:homelab-cluster |
| claude-haiku-4-5 | 424 | 66 | $0.000603 | ClawHub:homeserver |
| claude-haiku-4-5 | 405 | 74 | $0.000620 | ClawHub:pi-admin |
| claude-haiku-4-5 | 466 | 106 | $0.000797 | ClawHub:pi-health |
| claude-haiku-4-5 | 428 | 62 | $0.000590 | ClawHub:0x0-messenger |
| claude-haiku-4-5 | 409 | 78 | $0.000639 | ClawHub:trading |
| claude-haiku-4-5 | 413 | 79 | $0.000646 | ClawHub:trading-devbox |
| claude-haiku-4-5 | 414 | 76 | $0.000635 | ClawHub:trading-brain |
| claude-haiku-4-5 | 404 | 69 | $0.000599 | ClawHub:quant-trading-system |
| claude-haiku-4-5 | 416 | 58 | $0.000565 | ClawHub:openmm-grid-trading |
| claude-haiku-4-5 | 426 | 93 | $0.000713 | ClawHub:kalshi-cli-trading |
| claude-haiku-4-5 | 410 | 66 | $0.000592 | ClawHub:auto-trading-strategy |
| claude-haiku-4-5 | 419 | 65 | $0.000595 | ClawHub:futu-trading-bot |
| claude-haiku-4-5 | 426 | 66 | $0.000605 | ClawHub:trading-software-efficiency |
| claude-haiku-4-5 | 431 | 88 | $0.000697 | ClawHub:gate-exchange-trading-copilot |