Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Uplo Insurance
v1.0.0AI-powered insurance knowledge management. Search policy documents, claims records, underwriting guidelines, and actuarial data with structured extraction.
⭐ 0· 71·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The skill's name, description, README, SKILL.md, and identity patch all consistently describe an insurance knowledge search/GraphRAG capability (search_knowledge, search_with_context, get_directives, export_org_context). The declared capabilities align with the stated purpose. However, the registry summary at the top claims there are no required env vars/config, while skill.json requires an agentdocs_url and api_key — this mismatch is unexpected and reduces coherence.
Instruction Scope
SKILL.md limits runtime actions to MCP-style operations (get_identity_context, get_directives, search_knowledge, search_with_context, export_org_context, flag_outdated). It does not instruct access to unrelated local files or unrelated credentials. Note: the skill explicitly expects/encourages quoting sensitive policy and actuarial data from the org knowledge base, so it will surface confidential information if the connected UPLO instance contains it — this is consistent with its purpose but increases data-sensitivity risk.
Install Mechanism
The top-level metadata said 'No install spec' and 'instruction-only', but skill.json includes an mcp section that runs `npx -y @agentdocs1/mcp-server --http` and maps env to a configured AGENTDOCS_URL/API_KEY. README also shows using npx. That implies runtime fetching/execution of remote npm code via npx. Fetching/executing remote packages is a moderate-to-high risk operation unless you verify the package source and integrity. The inconsistency between 'no install' and an embedded npx command is also an implementation mismatch worth flagging.
Credentials
The only credentials/configuration referenced in skill.json and README are an UPLO instance URL and an API key (AGENTDOCS_URL / API_KEY), which are appropriate and proportional for a knowledge-base connector. However, the registry/requirements summary incorrectly lists 'none' for required env vars/config, which is inconsistent and could mislead users into not providing or reviewing credentials. No unrelated credentials are requested.
Persistence & Privilege
Flags show always:false and normal autonomous invocation settings. The skill does not request persistent system-wide privileges or modifications to other skills. There is no indication it will force-enable itself or persist beyond normal skill behavior.
What to consider before installing
This skill appears to do what it says (search your insurance knowledge base), but there are two issues to check before installing: (1) metadata mismatch — the registry shows no required config but skill.json and README require your UPLO/AgentDocs URL and an API key; verify and supply those intentionally, and (2) the skill will run `npx @agentdocs1/mcp-server` (pulling runtime code from npm). Before you proceed: verify the npm package (@agentdocs1/mcp-server) is an official/trusted package (review its npm page, source repo, and publisher), ensure the API key you provide has least-privilege access and appropriate audit logging, confirm the AGENTDOCS_URL points to your internal/approved UPLO instance (not a public or attacker-controlled endpoint), and consider restricting network access or running the MCP server in an isolated environment. If you cannot verify the npm package or do not want remote code pulled at runtime, ask the publisher for a vetted installation artifact or a signed release you can audit.Like a lobster shell, security has layers — review code before you run it.
latestvk979s9yffwhzsg8nhqw601qdsh8395m9
License
MIT-0
Free to use, modify, and redistribute. No attribution required.