Back to skill

Security audit

Temporal Knowledge Graph Synthesizer

Security checks for vulnerabilities and agentic risk

Overview

The skill’s knowledge-graph behavior is related to its purpose, but it automatically reads broad OpenClaw memory logs, writes persistent derived entity files, and includes an undisclosed optional cron setup that can keep running daily.

Install only if you are comfortable with the skill reading your OpenClaw daily memory logs and creating persistent entity summaries. Avoid running cron_setup.sh unless you explicitly want a daily background OpenClaw task, and review or remove generated files under ~/.openclaw/workspace/memory/entities/ if they contain sensitive names or projects.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • System PersistenceInstalls backdoors, hooks, services, or scheduled tasks that survive the run
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (3)

T06 · System Persistence

Error
Location
cron_setup.sh:5
Finding

Persistent Daily Agent Task Registration

Content
View full analysis
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
scripts/build_kg.py:14
Finding

Unbounded Processing and Persistent Derivation of Workspace Memory Data

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
SKILL.md:8
Finding

Unpinned Third-Party Packages and NLP Model Installation

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (6)

Tp4

High
Category
MCP Tool Poisoning
Confidence
92% confidence
Finding

The code partially matches the description because it does parse daily logs and extract entities into structured files. However, key parts of the declared purpose are not implemented. There is no actual temporal extraction from source content, no relationship extraction, and no graph construction beyond per-entity markdown documents. Additionally, the parser is narrowly scoped to date-named markdown files in one directory, not general unstructured memory files. Therefore the description materially overstates the implemented behavior.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding

The skill performs file-reading and file-writing behavior via an automatically executed background script, but it does not declare any explicit tool scope or permissions. This is dangerous because it obscures the skill's effective capabilities from users and policy layers, reducing informed consent and making unintended workspace data access or modification more likely.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The invocation language is broad enough that the skill may activate in situations users do not reasonably expect, especially because activation appears to trigger automatic parsing of recent memories. In a memory-processing skill, ambiguous triggering increases the chance of unintended access to sensitive workspace data and unanticipated file modifications.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill does not clearly warn that invocation causes automatic parsing of memory files and updates to generated artifacts. Hidden side effects on user data are a real security concern because they undermine informed consent and can expose or transform sensitive information without the user realizing it.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · scripts/build_kg.py (reported line 44)May include surrounding context.

python
else:
                    entities[name]["mentions"] += 1
    else:
        # Fallback simplistic extraction for demonstration
        words = re.findall(r'\b[A-Z][A-Za-z-]+\b', text)
        for w in words:
            if len(w) > 3:

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

The command downloads en_core_web_sm, which hard-codes an English NLP model. This imposes a language choice in the skill configuration without offering the user a locale option or explaining why English-only processing is required.

Content

No source excerpt is available for this finding.