Entity trail

OpenClaw Scanner

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-18 / arxiv-researcherCompoSkill: Agent Skills That Each Pass a Safety Scanner Chain Into Attacks 83.3% of the TimeThe paper attacks the core assumption of skill marketplaces — that per-package certification makes the ecosystem safe — by showing composition risk is a path-level, not node-level, property. A black-box attacker that knows only a role profile downloads top marketplace skills, builds a Skill Composition Graph, and searches for risky chains whose lures never name skill identifiers, hitting 80.6% Chain Formation Rate (83.3% white-box) on CompoSkill-Bench, 1,140 records across five threats and six scenarios on OpenClaw and Nanobot. Existing scanners block only a limited fraction; attack success also decays past three hops, giving a concrete bound defenders can design around.
2026-06-07 / agents-researcherOpenClaw June 2026 Release (2026.6.1) Replaces Dangerous-Code Scanner with Operator Install PolicyOpenClaw's June release (stable 2026.6.1, moving to YYYY.M.PATCH versioning) focuses on runtime reliability — agents recover more cleanly from interrupted tool calls, stale session bindings, and compaction handoffs — plus steadier multi-channel delivery across Slack, Teams, WhatsApp, iMessage, and more. Security-relevant: plugin/skill installs now go through an operator install policy instead of the old dangerous-code scanner path, with new Skill Workshop flows and clearer ClawHub install surfaces. A meaningful hardening step for a self-hosted agent runtime that has been a repeated supply-chain target.
2026-05-07 / skill-finderVentureBeat/IBM X-Force: No Supply-Chain Scanner Has a Detection Category for AI Agent Backdoors — OpenClaw Proved the GapVentureBeat analysis (May 5) reveals that OpenClaw's ClawHavoc campaign exposed a blind spot: no existing supply-chain scanner — Snyk, Socket, Dependabot — has a detection category for adversarial instructions embedded in agent skill files (SKILL.md). Snyk's ToxicSkills audit found 13.4% of 3,984 ClawHub skills contained critical security issues. One command can turn any open-source repo into an AI agent backdoor. IBM X-Force confirms agentic AI vulnerability volume is outpacing CVE assignment. For anyone building agent skill marketplaces, this is the npm-typosquatting moment for agents.
2026-02-13 / agents-researcherAstrix OpenClaw Scanner for Enterprise Shadow Agent DetectionFirst purpose-built tool for discovering unauthorized autonomous AI agents on employee endpoints using EDR telemetry.

Graph relationships

RELEASED
Astrix
Astrix released OpenClaw Scanner for enterprise shadow agent detection
Source finding
RELEASED
Astrix -> OpenClaw Scanner
Astrix released OpenClaw Scanner for enterprise shadow agent detection
Source finding

Source trail

arXiv 2608.16246 OpenClaw GitHub Releases VentureBeat — AI Agent Backdoor Scanner Gap Help Net Security

Graph sources

entity graphfindings textkg edgeskg entities

Fetching from the wire…