Back to skill

Security audit

Thoth Lite

Security checks for vulnerabilities and agentic risk

Overview

This README helper is not malicious, but it needs review because it installs an unpinned package into system Python and copies raw project file contents into the agent context.

Install only if you are comfortable with the selected project files being shown to the agent/model context. Use a narrow `PROJECT_PATH`, avoid folders containing secrets or private configuration, and prefer installing dependencies in a virtual environment with pinned versions instead of modifying system Python.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Findings (2)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:43
Finding

Unpinned Dependency Installation Bypasses System Package Protections

Content
View full analysis
Remediation
View remediation
" ``` 2. Pin `rich` and all transitive dependencies to reviewed versions. 3. Use a requirements or lock file containing cryptographic hashes, then install with `--require-hashes`. 4. Remove `--break-system-packages` so the command cannot bypass system package-management protections. 5. Remove `--quiet`, or otherwise retain installation logs for package source and version verification. 6. Configure an approved package index explicitly and require TLS certificate validation. 7. Consider replacing `rich` with standard-library output if formatted console rendering is not essential. ]]>

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:81
Finding

Recursive Project Scan Can Disclose Sensitive File Contents to the Agent Context

Content
View full analysis
Remediation
View remediation
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 (2)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill reads up to 20 files from a user-specified path and prints raw file contents directly into model-visible output without any privacy warning, consent gate, or filtering for secrets. In a documentation skill, this behavior can unintentionally disclose credentials, private source code, API keys, or sensitive configuration data if such data appears in scanned files.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill explicitly instructs dumping the first 3000 characters of each discovered file into the model context, which is a direct data exposure path. Because the target path can be a whole folder and the extension allowlist includes JSON, YAML, TOML, and Markdown, the skill may ingest secrets, internal docs, tokens, or personal data that are not necessary to generate a README.

Content

No source excerpt is available for this finding.