Back to skill

Security audit

Lab Units

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill benign with high confidence: the scanners found no embedded code, and the one flagged item is a disclosed installation command that carries supply-chain risk.

Before installing, consider reviewing the repository and using a verified, pinned version. Install in an environment without access to sensitive data or credentials. The verdict is benign, but it does not mean the installation is risk-free.

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
  • 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
Findings (1)

YARA rule 'agent_skill_remote_bootstrap_execution': Remote script or code download followed by execution/bootstrap installation [agent_skills]

High
Category
YARA Match
Confidence
87% confidence
Finding

The installation command downloads an unpinned Git repository and can execute its build or installation code with the invoking user's privileges. A compromised repository or malicious upstream change could therefore execute arbitrary code and potentially access sensitive laboratory CSVs available to that user. This is a conventional, explicitly disclosed installation step rather than evidence of malicious intent, but the tool's offline runtime does not mitigate installation-time supply-chain risk.

Content

Scanner excerpt · SKILL.md (reported line 23)May include surrounding context.

ion depends on the molecule, so a generic unit library cannot do it. Every shipped factor was verified against two independent authorities, and the factor used is written next to every converted column so each number can be checked by hand. Analytes with no fixed factor are refused, with the reason printed. Conversion only: no reference ranges, no flagging, no interpretation.

Install

bash
pip install git+https://github.com/RakTikoo/labunits

This provides the labunits command. labunits list shows the shipped analytes with units and factors; verify an install with labunits check (12 published worked pairs, PASS/FAIL).

Run

bash
labunits convert labs.csv --columns glucose,cholesterol_total,creatinine,hba1c,bun --from conventional --to si --out labs_si.csv
labunits convert-one --analyte glucose --value 100 --from conventional --to si
  • --columns takes comma-separated analyte keys; each key must also be a column in the file. If the file's column names diff