Back to skill

Security audit

Skill Build Wizard

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed workflow guide for building and publishing skills, with a minor path-discovery weakness but no evidence of hidden, destructive, or data-stealing behavior.

Install this if you want a guided skill-building and release workflow. Before publishing, verify the companion skills are trusted and only reply with the required publish confirmation after reviewing the change summary, acceptance report, version, and changelog. If multiple copies of this skill may exist, prefer running the pre-check script from the known installed path.

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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The declared description presents a broad orchestration skill that guides users through the full skill-building lifecycle and invokes multiple other tools for scaffolding, validation, regression, and release. The supplied code chunk only contains a pre-check shell script. Its behavior is limited to parsing skip flags, checking for a git repo in the current or common workspace paths, printing two recommendations, and returning a status summary. That pre-flight behavior is consistent with one small part of the declared workflow, but the overall declared purpose substantially overstates what this code chunk actually does. There are no obvious undeclared dangerous capabilities; the mismatch is that the implementation is materially narrower than the description.

Content

No source excerpt is available for this finding.

Agent Config Directory Access

High
Category
Agent Snooping
Confidence
86% confidence
Finding

The skill instructs the agent to search agent configuration/workspace directories under $HOME and /app to locate and execute a script. Accessing hidden agent directories can expose internal configuration layout and increases the chance of operating on unintended files, especially if a malicious or tampered skill directory with the same name is present in one of the searched locations.

Content

Scanner excerpt · SKILL.md (reported line 50)May include surrounding context.

bash
# Auto-locate the script (works under common locations)
SKILL_DIR=$(find "$HOME/.openclaw/workspace/skills" "$HOME/.claude/skills" /app/skills \
  -maxdepth 1 -name "skill-build-wizard" -type d 2>/dev/null | head -1)
bash "${SKILL_DIR}/scripts/pre-check.sh"

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 112)May include surrounding context.

md
See `references/checklist-coding.md` for the full checklist. Run four checks

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 197)May include surrounding context.

md
See `references/checklist-coding.md` for the full checklist. Run four checks

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 134)May include surrounding context.

md
- The skill directory should only contain release files: `SKILL.md`, `scripts/`,

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
85% confidence
Finding

The find command enumerates multiple skill-storage locations to discover a directory named skill-build-wizard. Enumeration of installed skill paths leaks environment structure and, combined with selecting the first match, can enable confused-deputy behavior or execution of an attacker-planted lookalike skill in a searched location.

Content

Scanner excerpt · SKILL.md (reported line 50)May include surrounding context.

bash
# Auto-locate the script (works under common locations)
SKILL_DIR=$(find "$HOME/.openclaw/workspace/skills" "$HOME/.claude/skills" /app/skills \
  -maxdepth 1 -name "skill-build-wizard" -type d 2>/dev/null | head -1)
bash "${SKILL_DIR}/scripts/pre-check.sh"