Real Skill packageSource verifiedClawHub registry

deepsafe-scan

Preflight security scanner for AI coding agents — scans deployment config, skills/MCP servers, memory/sessions, and AI agent config files (hooks injection) for secrets, PII, prompt injection, and dangerous patterns. Runs 4 model behavior probes (persuasion, sandbagging, deception, hallucination). Supports LLM-enhanced semantic analysis. Works with OpenClaw, Claude Code, Cursor, and Codex. Use when a user asks for a security audit, health check, or wants to scan their AI agent setup for vulnerabilities.

Identity and source

Publisher attributionXiaoYiWeioregistry owner unverified by skillvetai
Functional categoryAgent Engineering, Security & Governanceautomatically inferred · 83% rule confidence
Package forminstruction with code17 recorded files
Canonical sourceClawHub registryclawhub:xiaoyiweio:deepsafe-scan
Open canonical source ↗

Platform declarations

These states come from the source or distribution context. None of the entries below are SkillVetAI compatibility test results.

OpenClawnative officialProvenance: registry distribution

Independent structural checks

These checks parse the fixed package against dated platform rules. They do not execute the Skill or verify task behavior.

Claude Codepasses structure
Checker 0.1.0 · agent-skills-2026-08-13+claude-code-docs-2026-08-13 · 9/12/2026.claude/skills/deepsafe-scan

Runtime, accounts, dependencies, permissions, network behavior and task quality remain untested.

OpenAI Codexpasses structure
Checker 0.1.0 · agent-skills-2026-08-13+codex-docs-2026-08-13 · 9/12/2026.agents/skills/deepsafe-scan

Runtime, accounts, dependencies, permissions, network behavior and task quality remain untested.

OpenClawpasses structure
Checker 0.1.0 · agent-skills-2026-08-13+openclaw-docs-2026-08-13 · 9/12/2026skills/deepsafe-scan

Runtime, accounts, dependencies, permissions, network behavior and task quality remain untested.

Installation and inspection

This command is recorded from the source ecosystem and resolves the registry's latest release. The fixed release shown on this page should be inspected before adoption.

clawhub install @xiaoyiweio/deepsafe-scan
clawhub inspect @xiaoyiweio/deepsafe-scan --version 2.0.0

Security evidence

SkillVetAI static result: high signal

This automated, non-executing scan is bound to this release hash. It is not a safety certification and may contain false positives or false negatives.

Status
completed
Coverage
partial text content
Files
16 / 17 inspected as text
Checked
9/12/2026, 4:34:07 AM
Scanner
0.1.3
Policy
1.0.3
4 automated findings
highRemote content is piped directly to a shellscripts/scan.py:790 · confidence 98%curl|sh
highRemote content is piped directly to a shelldocs/plan-cross-platform-evolution.md:413 · confidence 98%curl|sh
highRemote content is piped directly to a shellREADME.md:156 · confidence 98%curl\|sh
highInstructions or code access sensitive credential locationsREADME.md:158 · confidence 85%~/.ssh/id_rsa
4 High/Critical review queue entries
STATIC_DOWNLOAD_PIPE_TO_SHELLpending
STATIC_DOWNLOAD_PIPE_TO_SHELLpending
STATIC_DOWNLOAD_PIPE_TO_SHELLpending
STATIC_SENSITIVE_CREDENTIAL_ACCESSpending
Open human review queue →
6 inferred permission indicators
  • shell execution — automatically inferred
  • network access — automatically inferred
  • filesystem read — automatically inferred
  • filesystem write — automatically inferred
  • credential access — automatically inferred
  • browser control — automatically inferred
5 dependency and API indicators
  • api: api.anthropic.com
  • api: api.openai.com
  • api: evil.com
  • api: github.com
  • api: img.shields.io
External clawhub result: suspicious

This is registry-supplied evidence for the recorded release, not an independent SkillVetAI scan. Check the canonical source for the full report, scanner versions, scope, and current moderation state.

Evidence checked
4/2/2026, 7:26:57 AM
Release binding
Matches this record
  • vt: clean
  • llm: suspicious

Recorded files

The catalog stores hashes and an inventory summary for change detection. It does not republish the package contents.

Package content hashsha256:9be4db09772d952917a679aac84d183442cc7e8c0cad6ba4d977dc8c9730d16f
Show up to 17 recorded paths
  • _meta.json
  • AGENTS.md
  • CLAUDE.md
  • data/deceptionbench.json
  • data/halueval_samples.json
  • data/manipulation_persuasion_topics.json
  • data/prompts.json
  • data/qa_evaluation_instruction.txt
  • docs/plan-cross-platform-evolution.md
  • README.md
  • scripts/llm_client.py
  • scripts/probes/deception_probe.py
  • scripts/probes/halueval_probe.py
  • scripts/probes/persuasion_probe.py
  • scripts/probes/sandbagging_probe.py
  • scripts/scan.py
  • SKILL.md

Source changelog

v2.0.0: Cross-platform universal AI agent security gateway. Supports OpenClaw, Claude Code, Cursor, and Codex. New hooks injection detection, multi-platform LLM auto-detection, externalized prompt templates.