Security audit
Security checks for vulnerabilities and agentic risk
This skill is a cloud media-analysis client with silent identity creation, local token storage, and several code/documentation mismatches that users should review before installing.
Install only if you are comfortable sending lawn images/videos or their URLs to the lifeemergence.com cloud service and having report history tied to a silently created or reused local identity. Treat it as a cloud account-backed analysis client, not a purely local lawn segmentation tool, and review or clear the workspace data database/tokens if you do not want reuse across runs.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
The combination of local file access, workspace detection, user/open-id generation, local DAO persistence, outbound API calls, and auth-token management represents a broad, stateful integration platform rather than a simple lawn-analysis skill. This mismatch is dangerous because it can silently create identities, persist user data, and communicate with remote services under a deceptively narrow description.
YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).
---
name: "smyx-lawn-health-assessment-analysis"
description: "AI-powered lawn health assessment from drone or fixed-camera top-down images. Uses semantic segmentation to distinguish healthy turf (green), wilting/yellow turf (yellow-brown), bare soil and weeds (off-species color/texture), then computes wilting area ratio and weed coverage ratio, and outputs a composite lawn health score (0-100). Helps managers of golf courses, courtyards or municipal greenways quantify turf quality and guide irrigation, fertilization and weeding operations. Scenarios: home courtyards, golf courses, municipal park lawns, sports fields. | 通过无人机或固
The file documents a pet health analysis API even though the skill is presented as a lawn health assessment capability. This kind of cross-domain mismatch is dangerous because it can cause the agent or integrator to invoke unrelated backend endpoints, expose the wrong data domain, or route user inputs and exported reports to an unintended service.
The endpoint paths and scene code explicitly reference pet health analysis, not lawn assessment, indicating the skill may be wired to the wrong service namespace. If used as-is, the skill could access or disclose unrelated pet-health workflows and records, creating data exposure, integrity, and authorization-boundary risks.
The manifest consistently describes lawn-health assessment from drone or fixed-camera top-down images, using image semantic segmentation. However, this code explicitly expects a local video path or network video URL and submits it as videoUrl, which broadens the skill from image analysis to video analysis in a way not reflected in the manifest.
The implementation materially contradicts the manifest: it performs generic video analysis and history listing rather than lawn-health analysis on top-down images. This kind of capability mismatch is dangerous because users and orchestrators may grant the skill data access, trust, or deployment approval based on the declared purpose while the code executes a different workflow, creating a supply-chain and deceptive-functionality risk.
The file contains broad identity management, token handling, local user persistence, and authenticated backend request capabilities that are unrelated to the stated lawn-health image analysis purpose. This unnecessary privileged functionality expands the attack surface and enables silent account creation and backend access from a skill users would reasonably expect to perform only image assessment.
The code reads a local workspace file for identity material and, if missing, auto-selects or creates a default user for remote service access. In the context of a lawn-image assessment skill, silently deriving identity from local state and provisioning remote identities is unjustified and can lead to unauthorized use of user/workspace identity, unexpected data association, and covert backend enrollment.
The skill declares broad operational behavior that implies shell execution, filesystem access, environment access, networking, and local persistence, but it does not define any explicit tool scope or least-privilege boundaries. That makes review and runtime enforcement harder and increases the chance that an agent can invoke powerful capabilities beyond what users expect from a lawn-image analysis skill.
Vague keyword-based activation for analysis can trigger processing on loosely related requests, especially when files or URLs are present. Because the skill also saves files and contacts remote APIs, accidental activation has privacy and data-handling consequences beyond mere inconvenience.
The skill expands from image analysis into cloud-hosted history retrieval and report-link generation, which changes the trust and privacy model. Users may not expect that requesting lawn analysis also exposes identity-linked remote records and clickable report URLs.
Overly broad trigger phrases for history lookup can cause the skill to invoke cloud report retrieval when the user did not intend it. In a skill with remote data access and identity linkage, ambiguous activation increases the risk of unneeded disclosure or unauthorized querying of historical records.
Automatically saving uploaded files locally without a clear user-facing warning is a privacy and retention risk. Users may assume transient processing, but local persistence can expose sensitive imagery to later access, backup systems, or other tools on the host.
Detected: suspicious.install_untrusted_source