Security audit
Security checks for vulnerabilities and agentic risk
The skill appears to perform the advertised cloud video analysis, but it also silently provisions or reuses an identity and persists authentication tokens locally, which deserves review before installation.
Install only if you are comfortable sending barn media and report queries to the lifeemergence.com/open.lifeemergence.com services and allowing the skill to create or reuse a local identity database with stored auth tokens. Review the local data retention and account-linking behavior before using it with sensitive farm footage or business records.
Outbound HTTP requests, auth-token/header handling, automatic user creation/login, local file I/O, database access, and workspace-context management all substantially exceed the stated behavior-detection purpose. This is dangerous because it combines remote account actions with local persistence and file access, creating a broad and opaque trust boundary that could expose data or perform unintended side effects.
Outbound HTTP requests, auth-token/header handling, automatic user creation/login, local file I/O, database access, and workspace-context management all substantially exceed the stated behavior-detection purpose. This is dangerous because it combines remote account actions with local persistence and file access, creating a broad and opaque trust boundary that could expose data or perform unintended side effects.
Outbound HTTP requests, auth-token/header handling, automatic user creation/login, local file I/O, database access, and workspace-context management all substantially exceed the stated behavior-detection purpose. This is dangerous because it combines remote account actions with local persistence and file access, creating a broad and opaque trust boundary that could expose data or perform unintended side effects.
Outbound HTTP requests, auth-token/header handling, automatic user creation/login, local file I/O, database access, and workspace-context management all substantially exceed the stated behavior-detection purpose. This is dangerous because it combines remote account actions with local persistence and file access, creating a broad and opaque trust boundary that could expose data or perform unintended side effects.
Outbound HTTP requests, auth-token/header handling, automatic user creation/login, local file I/O, database access, and workspace-context management all substantially exceed the stated behavior-detection purpose. This is dangerous because it combines remote account actions with local persistence and file access, creating a broad and opaque trust boundary that could expose data or perform unintended side effects.
Outbound HTTP requests, auth-token/header handling, automatic user creation/login, local file I/O, database access, and workspace-context management all substantially exceed the stated behavior-detection purpose. This is dangerous because it combines remote account actions with local persistence and file access, creating a broad and opaque trust boundary that could expose data or perform unintended side effects.
Outbound HTTP requests, auth-token/header handling, automatic user creation/login, local file I/O, database access, and workspace-context management all substantially exceed the stated behavior-detection purpose. This is dangerous because it combines remote account actions with local persistence and file access, creating a broad and opaque trust boundary that could expose data or perform unintended side effects.
Outbound HTTP requests, auth-token/header handling, automatic user creation/login, local file I/O, database access, and workspace-context management all substantially exceed the stated behavior-detection purpose. This is dangerous because it combines remote account actions with local persistence and file access, creating a broad and opaque trust boundary that could expose data or perform unintended side effects.
Outbound HTTP requests, auth-token/header handling, automatic user creation/login, local file I/O, database access, and workspace-context management all substantially exceed the stated behavior-detection purpose. This is dangerous because it combines remote account actions with local persistence and file access, creating a broad and opaque trust boundary that could expose data or perform unintended side effects.
Outbound HTTP requests, auth-token/header handling, automatic user creation/login, local file I/O, database access, and workspace-context management all substantially exceed the stated behavior-detection purpose. This is dangerous because it combines remote account actions with local persistence and file access, creating a broad and opaque trust boundary that could expose data or perform unintended side effects.
YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).
---
name: "smyx-aggressive-behavior-detection-analysis"
description: "Detects aggressive interactions in livestock and poultry from continuous barn videos — including fighting, biting, chasing and butting — and outputs behavior type, intensity level and alert level. | 识别打斗、撕咬等攻击行为，及时预警。"
version: "1.0.14"
license: "MIT-0"
---
# ⚔️ Livestock Aggressive Behavior Detection | 畜禽争斗行为识别
> 识别打斗、撕咬等攻击行为，及时预警。
>
> **群体冲突预警中枢** · 畜禽舍连续视频智能分析 · 争斗行为识别 · 强度/预警等级 · 历史报告云端查询
---
## �
The code provides remote modification primitives via generic add, edit, delete, POST, PUT, and DELETE wrappers without any visible scoping to behavior-analysis functions. In the context of a narrowly described analysis skill, unjustified write-capable network operations are dangerous because they could be abused to alter remote state, trigger unintended actions, or serve as a covert control channel if reachable by untrusted inputs.
The User model stores identity and authentication-related fields including username, realname, email, token, and open_token, but this is not justified by the skill's stated video-analysis function. Collecting and persisting such data without clear need materially raises privacy and credential-exposure risk, especially because the DAO supports broad read/write operations against a shared SQLite database.
When no identity is provided, the code generates a synthetic user identifier and persists a new user record locally for reuse. This is a strong capability escalation for a video-analysis skill because it creates long-lived identities and state without user approval, which can then be used to authenticate outbound requests and mask who initiated activity.
This utility layer performs identity provisioning, token acquisition, local user lookup, and external account login flows that are unrelated to livestock aggression detection. In the context of a video-analysis skill, this is dangerous because it silently expands scope into account creation and credential handling, enabling unintended external data transmission and backend access without a purpose tied to the advertised function.
The skill declares no explicit tool scope while its instructions require shell execution, local file handling, network access, and likely environment/context access. This is dangerous because it gives the agent broad, implicit capabilities without a manifest-level allowlist, making unintended or unsafe execution paths harder to constrain or audit.
The document is predominantly Chinese and specifies output conventions such as fixed Chinese column names and report names, but it does not indicate that users may choose another language. This can constitute a natural-language policy issue when a specific language is effectively enforced without user opt-in or documented locale justification.
The trigger conditions use broad behavioral and report-related keywords that can auto-invoke the skill in situations where the user did not intend to run it. This is dangerous because automatic execution may cause unnecessary file handling, network calls, or cloud history queries based on ambiguous language.
The manifest says this skill detects aggressive interactions in livestock and poultry from barn videos, but the exposed interface asks users to choose among 'cat', 'dog', and 'other'. That behavior suggests the implementation is repurposed from a generic pet-analysis flow rather than being scoped specifically to livestock/poultry as claimed.
Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
result_json = JsonUtil.parse(result_json_pure_text, result_json_pure_text)
result_json_common_ai_response = result_json.get("commonAiResponse") if isinstance(result_json,
dict) else result_json
if result_json_common_ai_response:
result_json = result_json_common_ai_response
Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
result_json = JsonUtil.parse(result_json_pure_text, result_json_pure_text)
result_json_common_ai_response = result_json.get("commonAiResponse") if isinstance(result_json,
dict) else result_json
if result_json_common_ai_response:
result_json = result_json_common_ai_response
Multiple user-facing strings are fixed in Chinese, such as the report headings and status/error messages, and the file provides no opt-in or locale selection mechanism. This can violate language/locale policy when a skill forces a specific language for all users without choice or documented justification.
The manifest describes detecting aggressive behavior from continuous barn videos, which suggests analyzing provided video content for that domain purpose. The code also accepts any HTTP/HTTPS URL as input and forwards it as videoUrl, enabling remote video retrieval from arbitrary sources rather than just local or clearly barn-scoped footage.
The skill reads arbitrary local files supplied by the user and uploads their full contents to an external analysis API without any in-code notice, confirmation, or narrowing beyond extension and size checks. This creates a privacy and data-handling risk because users may unintentionally exfiltrate sensitive local media or mislabeled files to a remote service.
This shared service exposes generic CRUD and arbitrary HTTP helper methods that are far broader than the declared purpose of analyzing aggressive livestock behavior from video. In a skill/plugin setting, this kind of overbroad network capability increases the attack surface and can be repurposed to access, query, or manipulate unrelated backend resources if other components pass attacker-controlled URLs or parameters.
Detected: suspicious.install_untrusted_source