Security audit

AC-Nexus

Security checks for vulnerabilities and agentic risk

Overview

This AC-control skill is coherent with its purpose, but it asks users to install mutable remote code and persist sensitive device credentials and autonomous schedules.

Install only after reviewing the referenced GitHub repository and dependencies, preferably from a pinned commit or signed release in a virtual environment under a non-admin account. Treat Xiaomi tokens, API keys, device names, location, and schedules as sensitive; protect ~/.ac_controller/config.json with owner-only permissions or a credential manager, and explicitly approve any persistent schedules, auto-adjustment, or storm shutdown behavior.

A.I.G

Vulnerability Patterns

Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally

Findings (2)

T03 · Remote Payload Retrieval and Execution

Warning

Location: SKILL.md:25
Finding: Mutable Remote Repository Is Retrieved and Installed Without Integrity Pinning
Content: View full analysis
Remediation: View remediation
``` 2. Publish and verify a SHA-256 checksum or signed attestation for the release artifact. 3. Pin every Python dependency to an exact version and include cryptographic hashes. 4. Install dependencies with hash enforcement: ```bash python -m pip install --require-hashes -r requirements-core.txt ``` 5. Include the implementation and dependency lock file in the auditable Skill package rather than delegating execution to mutable external content. 6. Use a dedicated virtual environment and a non-administrative account. 7. Add automated dependency provenance, vulnerability, and signature checks to the release process. ]]>

T09 · Insecure Skill Coding Practices

Warning

Location: SKILL.md:124
Finding: Xiaomi Device-Control Tokens Are Persisted in a Plain Configuration File
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (1)

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 87% confidence
Finding: The guidance says device names should be human-readable examples like Chinese room names and states that the agent matches user intent using a Chinese example phrase. This implies a language/locale expectation for successful use, but the document does not offer the user a language choice or clearly justify a Chinese-only constraint.
Content: No source excerpt is available for this finding.