Security audit

Edvisage Agent Ops

Security checks for vulnerabilities and agentic risk

Overview

This is an instruction-only monitoring skill that openly asks the agent to keep cost and activity logs, but users should treat those logs as potentially sensitive.

Install only if you are comfortable with the agent keeping persistent operational logs. Keep entries high-level, redact secrets and private targets, periodically clear old logs, and do not rely on this free version for real spending-limit enforcement.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
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 (3)

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The README promotes logging every action, tracking costs, and generating daily summaries, but it does not mention safeguards for sensitive data, retention limits, redaction, or user consent. In an agent context, comprehensive logs can easily capture prompts, tool inputs, outputs, secrets, personal data, or internal business information, creating avoidable privacy and compliance risk if operators enable the skill without understanding those implications.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill directs the agent to maintain detailed daily cost and activity logs in memory files, but it does not instruct the agent to obtain user consent, minimize logged data, or avoid capturing sensitive content. In practice, broad operational logging can accumulate prompts, targets, file paths, API usage patterns, and other metadata that may expose private user activity or internal system behavior if memory is later accessed, reused, or leaked.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The action logging guidance explicitly includes file reads, file writes, skill use, tool use, and API calls, which can create a detailed audit trail of sensitive operations without any warning about privacy, secret handling, or system-integrity risks. This increases the chance that confidential filenames, endpoints, internal workflows, or user-related activity are persistently recorded and later exposed to other skills, users, or compromised components.

Static analysis

No suspicious patterns detected.