Security audit
Security checks for vulnerabilities and agentic risk
The skill is a cloud pet-media analysis client, but it silently creates and reuses a persistent user identity and token store while uploading media and querying report history, so it needs user review before installation.
Install only if you are comfortable with pet images or videos being sent to lifeemergence.com/open.lifeemergence.com services, linked to an automatically managed local/remote identity, and having report history available later. Review or disable it if you need explicit consent before uploads, do not want local token persistence, or require clear retention/deletion controls for reports and metadata.
Authenticated API communication combined with local file access, workspace detection, local identity creation, and token/session management is a much broader capability set than the skill description suggests. In context, that means uploaded media may become linked to persistent identities and backend sessions without clear user understanding, increasing privacy and account-scope risk.
Authenticated API communication combined with local file access, workspace detection, local identity creation, and token/session management is a much broader capability set than the skill description suggests. In context, that means uploaded media may become linked to persistent identities and backend sessions without clear user understanding, increasing privacy and account-scope risk.
Authenticated API communication combined with local file access, workspace detection, local identity creation, and token/session management is a much broader capability set than the skill description suggests. In context, that means uploaded media may become linked to persistent identities and backend sessions without clear user understanding, increasing privacy and account-scope risk.
Authenticated API communication combined with local file access, workspace detection, local identity creation, and token/session management is a much broader capability set than the skill description suggests. In context, that means uploaded media may become linked to persistent identities and backend sessions without clear user understanding, increasing privacy and account-scope risk.
Authenticated API communication combined with local file access, workspace detection, local identity creation, and token/session management is a much broader capability set than the skill description suggests. In context, that means uploaded media may become linked to persistent identities and backend sessions without clear user understanding, increasing privacy and account-scope risk.
Authenticated API communication combined with local file access, workspace detection, local identity creation, and token/session management is a much broader capability set than the skill description suggests. In context, that means uploaded media may become linked to persistent identities and backend sessions without clear user understanding, increasing privacy and account-scope risk.
Authenticated API communication combined with local file access, workspace detection, local identity creation, and token/session management is a much broader capability set than the skill description suggests. In context, that means uploaded media may become linked to persistent identities and backend sessions without clear user understanding, increasing privacy and account-scope risk.
Authenticated API communication combined with local file access, workspace detection, local identity creation, and token/session management is a much broader capability set than the skill description suggests. In context, that means uploaded media may become linked to persistent identities and backend sessions without clear user understanding, increasing privacy and account-scope risk.
Authenticated API communication combined with local file access, workspace detection, local identity creation, and token/session management is a much broader capability set than the skill description suggests. In context, that means uploaded media may become linked to persistent identities and backend sessions without clear user understanding, increasing privacy and account-scope risk.
Authenticated API communication combined with local file access, workspace detection, local identity creation, and token/session management is a much broader capability set than the skill description suggests. In context, that means uploaded media may become linked to persistent identities and backend sessions without clear user understanding, increasing privacy and account-scope risk.
Authenticated API communication combined with local file access, workspace detection, local identity creation, and token/session management is a much broader capability set than the skill description suggests. In context, that means uploaded media may become linked to persistent identities and backend sessions without clear user understanding, increasing privacy and account-scope risk.
The manifest describes a skill for identifying cat/dog breeds and distinguishing individual pets, but this file handles arbitrary video input, submits it for a generic 'analysis', polls for report completion, and lists prior analysis reports. There is no pet-specific logic, breed recognition, or individual-animal identification behavior visible here; the outputs are framed as generic analysis reports.
The code implements generic video analysis and history retrieval rather than pet breed or individual pet recognition as advertised. This skill/implementation mismatch is dangerous because users or higher-level agents may grant inputs, permissions, or trust based on the declared pet-recognition purpose while the code performs different remote-analysis actions, creating a capability-deception and data-misrouting risk.
This file exposes a generic API wrapper with broad CRUD methods and arbitrary HTTP helpers that are not scoped to the advertised pet breed/individual recognition purpose. In a skill context, such overly broad network capability can be repurposed to access unrelated backend functionality, move data between services, or act as a general proxy, which significantly expands attack surface beyond the declared feature set.
The http_post/http_put/http_get/http_delete methods accept caller-controlled URLs and directly perform remote requests, creating an arbitrary outbound request primitive. In a skill unrelated to generic networking, this is especially dangerous because it can enable SSRF-style access to internal services, exfiltration to attacker-controlled hosts, or abuse of the host environment as a network pivot.
This file provides generic DAO logic plus user/account persistence for usernames, real names, email, tokens, and deletion state, which is unrelated to a pet breed/individual recognition skill. Capability mismatch is dangerous because it creates hidden identity and account-management functionality that can collect, mutate, or persist user data outside the stated purpose, increasing the chance of covert tracking or unauthorized retention.
The User model stores identity and authentication-related fields such as username, realname, email, token, and open_token without any evident need for pet recognition. In this context, token storage is especially sensitive because compromise or misuse could enable account correlation, impersonation, or long-term tracking far beyond the advertised function of identifying pets.
This utility code performs identity discovery, default account generation, local persistence, and reads credential-related data from the workspace, which is unrelated to a pet recognition skill's stated purpose. The hidden creation and reuse of user identities increases the risk of undisclosed tracking, unauthorized account binding, and unintended use of local credentials.
The HTTP wrapper silently authenticates, auto-registers users, persists tokens, and injects identity fields into outbound requests. For a pet breed/individual recognition skill, this is materially over-scoped behavior that can transmit user-linked data to external services without meaningful transparency or opt-in.
The skill advertises and requires powerful capabilities including shell execution, local file handling, environment access, and network access, but it declares no explicit tool scope or permission boundaries. In an agent environment, this increases the chance of unintended or over-broad execution paths, especially because the document contains imperative instructions to save files locally and invoke scripts against remote services.
An ambiguous default activation condition makes it unclear when the skill is allowed to handle user inputs and media, which can lead to unintended file saving, network submission, or analysis. In security-sensitive agent contexts, vague activation boundaries increase the risk of accidental overreach and data exposure.
Overly broad trigger phrases for history-report lookup can cause the skill to invoke cloud report-listing functionality during ordinary conversation without clear user intent. In a skill that links reports to internal identities, accidental triggering can expose sensitive historical metadata or cause unnecessary network access.
The skill mandates internal identity handling and even automatic creation or reuse of a default local user, which is unrelated to the core task of identifying pets in media. This is dangerous because it silently introduces persistent user linkage and local account state, enabling cross-session tracking or accidental exposure of report history.
The privacy statement says uploaded video data is deleted immediately and not persistently stored, yet other sections describe cloud-hosted historical report listing and report links. This inconsistency is dangerous because users may rely on false retention assurances while derived data or report artifacts remain stored and accessible.
This code defines the command description and multiple help strings entirely in Chinese, and later prints Chinese status/error messages as well. Because the file provides no opt-in language selection or documented regional justification, it imposes a specific language on users, which matches the natural-language locale policy violation criteria.
Detected: suspicious.install_untrusted_source
