ClawHub
Back to skill

Security audit

BioSkills

Security checks for vulnerabilities and agentic risk

Overview

This is a legitimate-looking bioinformatics skill installer, but it gives one command broad, persistent authority to install hundreds of agent skills and execute a secondary script from a cloned repository.

Install only if you trust the GPTomics bioSkills repository at the pinned commit and are comfortable adding many persistent skills to your OpenClaw environment. Prefer using --categories for the specific bioinformatics areas you need, review the cloned install-openclaw.sh before running, and be careful with --uninstall because it removes all bio-* skill directories.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (6)

Self-Modification

High
Category
Rogue Agent
Confidence
85% confidence
Finding

Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Content

Scanner excerpt · scripts/install-bioskills.sh (reported line 37)May include surrounding context.

sh
echo ""
    echo "Options:"
    echo "  --categories CATS     Install only specified categories (comma-separated)"
    echo "  --update              Only update skills that have changed"
    echo "  --uninstall           Remove all bio-* prefixed skills"
    echo "  --verbose             Show detailed output"
    echo "  --help                Show this help message"

Chaining Abuse

High
Category
Tool Misuse
Confidence
75% confidence
Finding

Tool calls are chained to bypass individual safety checks or escalate capabilities beyond what any single tool call would allow.

Content

Scanner excerpt · scripts/install-bioskills.sh (reported line 77)May include surrounding context.

sh
fi
    done
    echo "Removed $removed skills."
    [ -d "$INSTALL_DIR" ] && rm -rf "$INSTALL_DIR" && echo "Removed cached repository."
    exit 0
fi

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill instructs users to run a bundled shell script that downloads and installs a very large collection of external content, but it provides no warning about network access, filesystem changes, dependency installation, or trust boundaries. In a meta-skill that pulls 425 additional skills from a remote source, this creates significant supply-chain and execution risk because users are asked to execute opaque installer logic without review or pinning.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The statement that installed skills are 'automatically triggered based on the task at hand' defines activation in overly broad terms and does not establish clear boundaries for when specialized bioinformatics skills should run. In a package that installs hundreds of skills, ambiguous triggering increases the chance of unintended execution, surprising tool use, or inappropriate handling of sensitive biological or clinical data.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The script exposes an uninstall mode that performs recursive deletion of all directories matching $HOME/.openclaw/skills/bio-*. Although documented, this is still destructive behavior beyond simple installation and could remove legitimate user content if unrelated skills or directories share that prefix. In the context of a skill installer, filesystem deletion increases risk because users may invoke it expecting only install/update behavior.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

After cloning a remote repository, the installer executes a repository-supplied shell script (install-openclaw.sh), which grants that remote content arbitrary code execution in the user's environment. The pinned commit reduces supply-chain risk, but the behavior is still dangerous because the top-level installer delegates trust to a second script whose full actions are not constrained by this wrapper.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.