ClawHub
Back to skill

Security audit

architect-project-state

Security checks for vulnerabilities and agentic risk

Overview

This skill is a scoped document-management aid for architectural project state, with clear limits on reading, writing, versioning, and human confirmation.

Use a narrow project root and review generated state files before relying on them. The skill is designed to preserve provenance and avoid treating proposals or meeting notes as approvals, but it is not a substitute for client, regulatory, or professional sign-off.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (2)

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 159)May include surrounding context.

md
- Any claimed update has actually read its predecessor; otherwise the artifact is labelled reconstruction or comparison.
- No unconfirmed meeting item is labelled decided, fixed, approved, rejected, or final.
- “Not recommended for large-area use” has not become “material rejected”.
- “No approval to sacrifice area” has not become “area loss prohibited”.
- No prior Agent proposal was inherited as project fact.
- Every numeric comparison uses compatible definitions or states why comparison is blocked.
- Saleable or marketing area has not been declared to exceed or close against FAR-counted area without an explicit mapping source.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

Line L077 gives special handling instructions for specific Chinese phrases, which introduces language-specific behavior in the skill's natural-language rules. The file does not explicitly state that the skill is limited to Chinese-language source materials or that this language-specific rule is optional or user-selected.

Content

No source excerpt is available for this finding.