Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed Markdown-only analysis aid for comparing conflicting viewpoints, with no hidden execution, credential use, persistence, or broad data access found.
Install this if you want OpenClaw to structure complex research around competing claims and source bias. Be aware it may activate for broad debate-style questions and make answers more conflict-focused; review the uninstall path before running deletion commands.
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
rmdir /S /Q %USERPROFILE%.openclaw\skills\contradiction-analyzer
rm -rf ~/.openclaw/skills/contradiction-analyzer
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
rmdir /S /Q %USERPROFILE%.openclaw\skills\contradiction-analyzer
rm -rf ~/.openclaw/skills/contradiction-analyzer
The skill declares automatic trigger conditions using very broad, common conversational patterns such as open-ended discussions, controversial topics, and generic phrases like 'different views' or 'pros and cons'. This can cause the skill to activate unintentionally during ordinary user interactions, expanding its influence beyond user intent and potentially altering responses, source weighting, or analysis framing without explicit consent.
The skill says to 'automatically activate' for broad categories like 'open-ended discussions,' 'controversial topics,' and common keywords such as 'opinions' or 'pros and cons.' These conditions overlap with ordinary conversation and do not provide clear boundaries or negative examples beyond a single exclusion for simple factual questions.
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.
## 3. Extreme Viewpoint Identification
⚠️ **Flagged:**
- "UBI will solve all poverty" - Overly simplistic, ignores structural factors
- "UBI will destroy the work ethic" - Contradicts pilot evidence, overly pessimistic
- "UBI is impossible to fund" - Ignores potential revenue sources, assumes status quo
The trigger conditions are broad enough to activate on many ordinary analytical queries, including generic open-ended or multi-factor questions. This can cause unintended invocation of the skill, steering responses into contradiction framing even when the user did not request it, which may distort outputs, overemphasize conflict, or interfere with higher-priority instructions.
The uninstall section provides destructive deletion commands but does not explicitly warn users that the removal is permanent and should be reviewed before execution. While the path shown is narrowly scoped to the skill directory, copy-pasting destructive commands without cautionary guidance increases the risk of accidental data loss if the path is modified or expanded.
The instructions include Chinese text fragments ("贯穿始终" and "倾向") within otherwise English guidance, which imposes mixed-language output or interpretation without user opt-in. The file does not state that multilingual output is optional, user-selectable, or required for a specific regional purpose.
No suspicious patterns detected.