Security audit

DeepLink Agentic

Security checks for vulnerabilities and agentic risk

Overview

This skill mostly matches its real-estate research purpose, but it exposes renewed access tokens in cleartext and includes public-sharing and deletion actions without strong safeguards.

Install only if you are comfortable sending prompts and selected files to agentic.dichanai.com. Treat AGENTIC_TOKEN and any NEW_TOKEN output as secrets, avoid running renewal in shared or logged terminals, and require explicit user confirmation before sharing, scheduling, or deleting tasks.

SkillSpector

Vulnerability Patterns

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

Findings (4)

Context-Inappropriate Capability

Medium

Confidence: 96% confidence
Finding: The skill instructs the agent to handle token renewal and states that a new token may be printed to the terminal so the agent can update the environment. Any workflow that surfaces fresh authentication tokens to the agent or logs materially increases the risk of credential leakage, reuse, or exfiltration beyond the user's intended real-estate task.

Context-Inappropriate Capability

Medium

Confidence: 85% confidence
Finding: The skill exposes a public sharing capability for tasks, which can make research content, uploaded files, or generated outputs accessible beyond the original user context. In a workflow that may include proprietary market analysis or private documents, public sharing is a meaningful confidentiality risk unless tightly constrained and explicitly consented to.

Missing User Warnings

Medium

Confidence: 98% confidence
Finding: The code prints a freshly issued bearer token directly to stdout, which can expose credentials to shell history, terminal logs, CI logs, screen recordings, or other local observers. Because this token grants authenticated access to the remote service, disclosure can lead to account or data compromise within the token's scope.

Missing User Warnings

Medium

Confidence: 98% confidence
Finding: Manual renewal also emits the complete new bearer token without masking or warning, creating the same secret-exposure risk through logs, terminal capture, and process output collection. In operational environments this is especially risky because CLI output is often archived or copied into tickets and chat systems.

Static analysis

No suspicious patterns detected.