Back to skill

Security audit

Edithai

Security checks for vulnerabilities and agentic risk

Overview

EdithAI is purpose-aligned as a log-analysis skill, but it asks users to install an unpinned global third-party CLI that will handle sensitive logs, an API key, file access, and diagnostic commands.

Install only if you are comfortable trusting the external npm CLI and DeepSeek with the logs you choose to analyze. Prefer a pinned and reviewed CLI version, avoid running installation or analysis with administrative privileges, redact secrets before analysis, use a narrowly scoped API key, and review or clear `~/.edithai/history.json` when working with sensitive incidents.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T08 · Insecure Dependencies

Error
Location
SKILL.md:45
Finding

Unpinned Global Installation of an Unaudited Third-Party CLI Package

Content
View full analysis
Remediation
View remediation
``` The selected version must be independently reviewed before publication of the skill. 2. **Verify package integrity** Record and verify the expected npm integrity digest or package tarball checksum. Automate verification in the installation process so installation fails if package contents change. 3. **Vendor or include the implementation** Include the executable source and dependency manifests in the audited artifact, or reference an immutable source commit and reproducible build instructions. This permits review of command execution, file access, API communication, and security controls. 4. **Avoid global installation** Prefer a project-local dependency with a committed lockfile, or execute the tool from a dedicated isolated environment. Do not recommend `sudo npm install -g`. 5. **Control lifecycle scripts** Use `--ignore-scripts` where package functionality permits. If lifecycle scripts are required, audit each script and all code it invokes before installation. 6. **Lock transitive dependencies** Commit a lockfile with integrity metadata and use a deterministic installation command such as `npm ci`. Continuously scan direct and transitive dependencies for compromise and known vulnerabilities. 7. **Apply runtime isolation** Run the CLI in a container or sandbox with: - Read-only access to explicitly selected log files. - No access to unrelated home-directory or system files. - A minimal environment containing only required credentials. - Restricted outbound network access limited to the expected API endpoint. - No administrative privileges. - Command execution disabled unless explicitly necessary. 8. **Reduce credential exposure** Use a narrowly scoped API ...[truncated 423 chars]
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (11)

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · CAPABILITIES.md (reported line 94)May include surrounding context.

edithai -query "show unusual file access patterns outside business hours"

Privilege escalation detection

edithai -query "detect any sudo or elevation attempts in auth logs"

text

### 3. Performance Optimization

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The guide instructs users to configure a DeepSeek API key and use a remote AI service, but it does not warn that prompts, log contents, or potentially sensitive file data may be transmitted to an external provider. In a log-analysis context, this is significant because logs often contain secrets, tokens, internal hostnames, stack traces, and personal data.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · INSTALL.md (reported line 49)May include surrounding context.

Configuration

Optional Configuration File

Create a configuration file at ~/.edithai/config.json for custom settings:

json
{

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The manifest describes an intelligent log analysis CLI with file operations, system diagnostics, and log pattern recognition, but this installation guide explicitly introduces configurable command whitelists/blacklists for terminal operations. Executing arbitrary terminal commands is a distinct capability not clearly justified by log analysis itself, and the documentation indicates this is a supported feature rather than an incidental implementation detail.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

This markdown file states that conversation history is stored at ~/.edithai/history.json, but it does not warn users that their prompts or analysis context may be retained on disk. Because stored history can contain sensitive operational or log data, the documentation should disclose this behavior clearly.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The README promotes AI-powered log analysis via the DeepSeek API but does not clearly warn users that log contents may be transmitted to a third-party service for processing. Logs often contain credentials, tokens, internal hostnames, stack traces, and personal data, so omission of this disclosure can lead users to exfiltrate sensitive information unintentionally.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill prominently advertises DeepSeek-powered log analysis but does not clearly warn users that log contents may be sent off-host to a third-party API. Because logs often contain secrets, tokens, internal hostnames, customer data, or incident details, this omission can cause unintended data exfiltration and privacy/compliance violations.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
80% confidence
Finding

The interactive, multi-turn mode combined with a documented history file indicates session persistence of prompts and analysis context. In a log-analysis tool, retained conversation history can store sensitive log excerpts, investigative notes, and credentials copied from logs, increasing exposure on disk and across later sessions.

Content

Scanner excerpt · SKILL.md (reported line 46)May include surrounding context.

edithai -query "find all database connection errors"

Generate performance report

edithai -query "create performance summary from application logs"

text

### Interactive Mode

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
77% confidence
Finding

This markdown file advertises "Write Reports" as a capability, which implies file creation or modification. While the document lists capabilities, it does not clearly warn users that the skill may write output files to disk or affect local data.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
67% confidence
Finding

The file states that the skill can perform network analysis and includes "DeepSeek cost estimation integration," which may involve external or internal network communication. The description does not disclose whether any user or system data could be transmitted during these operations.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

The skill examples and configuration imply report generation and persistent history storage, but the description does not clearly warn that the tool can write local files and retain session artifacts. This can surprise users in sensitive environments where writing reports or histories to disk may leak operational data or violate handling requirements.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.