Security audit
Security checks for vulnerabilities and agentic risk
This is a read-only brand intelligence plugin that sends brand lookup inputs to a configured API, with no hidden local access, persistence, or destructive behavior found.
Install this only if you are comfortable sending brand search terms, company slugs, capability queries, and watchlist brand lists to geo.sig.ai or to the configured baseUrl. Avoid using sensitive internal project names as lookup terms unless that disclosure is acceptable, and prefer reviewed dependency versions in managed deployments.
The activation guidance is overly broad because it triggers when a user merely mentions a company and the agent 'needs background context,' not only when the user explicitly requests a brand brief. This can cause unnecessary tool invocation, unintended disclosure of external-source content, or off-task behavior by injecting brand summaries into conversations where they were not requested.
The trigger conditions are broad enough to match many ordinary comparison requests such as 'X vs Y' or 'which is better,' which can cause this skill to activate outside a narrowly scoped brand-comparison workflow. This is dangerous because overbroad routing can misapply the skill, invoke comparison-specific tools when not appropriate, and reduce control over which capabilities are exposed in general conversation.
The invocation guidance includes phrases like "what's happening with my companies?" and "update me on X, Y, Z," which are broad natural-language requests that could plausibly occur in ordinary conversation. The file does not provide tighter constraints, exclusion conditions, or negative examples clarifying when this skill should or should not activate.
The dependency uses a caret range (^0.34.0) instead of an exact version, which makes builds non-reproducible and can pull in newer releases without explicit review. In a plugin/agent ecosystem, this increases supply-chain risk because a compromised or breaking upstream release could be consumed automatically.
"skills"
],
"dependencies": {
"@sinclair/typebox": "^0.34.0"
},
"peerDependencies": {
"openclaw": ">=2026.3.24-beta.2"
The peer dependency on openclaw is expressed as a broad minimum version (>=2026.3.24-beta.2), so consumers may install any compatible version, including versions affected by known advisories. Because the manifest does not constrain to a known-fixed range, it is impossible to verify from this file alone whether deployments are protected.
This code performs outbound network requests and may send user-provided values such as search queries, slugs, capabilities, and digest inputs to a remote API. Within this file there is no confirmation prompt, logging, comment, or docstring disclosing that user-supplied data will be transmitted off-system.
This plugin registers multiple tools that send user-supplied parameters to an external service via BrandApiClient, but the file provides no user-facing disclosure that queries and brand lookups leave the local agent environment. In an agent/tooling context, undisclosed external transmission can expose sensitive prompts, business interests, or internal identifiers to a third-party API, even if the behavior is functionally intended.
No suspicious patterns detected.