中文 EN

Scan Report

Scan New Files

⚠ Low Risk — Risk Score 20/100
Last scan:19 hr ago Rescan
20 /100
uplo-architecture
AI-powered architecture knowledge management. Search building designs, code compliance records, project specifications, and BIM data with structured extraction.
Legitimate MCP-based knowledge management skill with proper credential handling, though the supply chain dependency on @agentdocs1/mcp-server via npx introduces moderate third-party execution risk.
Skill Nameuplo-architecture
Duration70.6s
Enginepi
✓
Safe to install
Approve for use with standard sandboxing. Consider pinning @agentdocs1/mcp-server to a specific version in production to reduce supply chain risk.

Findings 2 items

Severity Finding Location
Low
Unpinned npm dependency via npx Supply Chain
The skill executes @agentdocs1/mcp-server via 'npx -y' without version pinning. This allows any version to be downloaded at execution time, potentially introducing malicious code if the package is compromised.
"command": "npx", "args": ["-y", "@agentdocs1/mcp-server", "--http"]
→ Pin the dependency to a specific version: "@agentdocs1/[email protected]" to ensure reproducible and audited behavior.
 skill.json:19
Low
Documentation capability mismatch Doc Mismatch
SKILL.md, skill.json, and README.md list different capabilities. SKILL.md shows 5 tools, skill.json shows 5 different tools, README.md references 21 MCP tools. This inconsistency could hide additional functionality.
SKILL.md lists: search_knowledge, search_with_context, export_org_context, get_directives, log_conversation
→ Unify capability documentation across SKILL.md, skill.json, and README.md to ensure transparency.
 SKILL.md:1
ResourceDeclaredInferredStatusEvidence
Network READ READ ✓ Aligned MCP server makes HTTP requests to configured agentdocs_url
Environment NONE READ ✓ Aligned API_KEY passed to external MCP server as configured secret
Shell NONE WRITE ✓ Aligned npx execution is documented and necessary for MCP server startup
10 findings
🔗
Medium External URL 外部 URL
https://img.shields.io/badge/ClawHub-uplo-architecture-blue
README.md:5
🔗
Medium External URL 外部 URL
https://clawhub.com/skills/uplo-architecture
README.md:5
🔗
Medium External URL 外部 URL
https://img.shields.io/badge/MCP-21_tools-green
README.md:6
🔗
Medium External URL 外部 URL
https://img.shields.io/badge/schemas-4-orange
README.md:7
🔗
Medium External URL 外部 URL
https://uplo.ai/schemas
README.md:7
🔗
Medium External URL 外部 URL
https://your-instance.uplo.ai
README.md:24
🔗
Medium External URL 外部 URL
https://clawhub.com/skills/uplo-engineering
README.md:60
🔗
Medium External URL 外部 URL
https://clawhub.com/skills/uplo-knowledge-management
README.md:61
🔗
Medium External URL 外部 URL
https://clawhub.com/skills/uplo-accounting
README.md:62
🔗
Medium External URL 外部 URL
https://app.uplo.ai
skill.json:17

File Tree

4 files · 7.1 KB · 185 lines
Markdown 3f · 136L JSON 1f · 49L
├─ 📝 identity-patch.md Markdown 9L · 1.7 KB
├─ 📝 README.md Markdown 70L · 2.7 KB
├─ 📋 skill.json JSON 49L · 1.2 KB
└─ 📝 SKILL.md Markdown 57L · 1.5 KB

Dependencies 1 items

PackageVersionSourceKnown VulnsNotes
@agentdocs1/mcp-server latest npm (npx) No Version not pinned - downloaded via npx -y at runtime

Security Positives

✓ No local code execution or script files present - purely a wrapper around external MCP server
✓ API key properly marked as secret in skill.json configuration
✓ No filesystem access declared or observed
✓ No credential harvesting or exfiltration patterns detected
✓ No obfuscation or base64-encoded execution
✓ No sensitive path access (ssh, aws, .env) observed
✓ MCP transport is documented and uses standard HTTP protocol