Security audit
Security checks for vulnerabilities and agentic risk
The skill is mainly a Buddhist study helper, but it includes unrelated trading and monetization guidance that users would not expect from that purpose.
Install only if you are comfortable with a Buddhist-learning skill that stores learned summaries locally and may use web search. Review or remove the unrelated learned files about weather trading, OpenClaw monetization, and psychedelic activation before trusting the skill for normal Dharma study.
Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.
#### 3. **Evening Rest (10 min)**
**Practice:**
- Review the day (without judgment)
- Recognize: "All experiences were like dreams"
- Rest in natural state (5 min)
- Sleep meditation: "May I recognize the clear light of sleep"
This file is clearly out of scope for a Buddhist learning skill and instead contains a detailed monetization and trading-intelligence plan centered on prediction-market arbitrage. Scope drift like this is dangerous because it can repurpose the agent into providing regulated financial/gambling-adjacent guidance and autonomous business automation that users would not reasonably expect from the declared skill.
The described capabilities include automated monitoring, mismatch detection, edge calculation, and alerting for weather-based trading opportunities, which amounts to operational trading intelligence. In the context of a Buddhist learning skill, these unjustified capabilities materially increase misuse risk by enabling financial decision support, possible circumvention of gambling restrictions through 'analysis-only' packaging, and deceptive expansion of the agent's effective permissions and purpose.
The README advertises searching authoritative web sources but does not warn users that their queries or topic interests may trigger outbound network requests. This is dangerous because prompts may contain sensitive personal, spiritual, or identifying information that could be exposed to third-party services or logged externally without the user's informed consent.
The README states that learned summaries are automatically saved to disk, but it does not clearly warn users that their prompts, derived summaries, or potentially sensitive religious/practice notes may be persisted locally. This creates a privacy and data-handling risk because users may disclose personal beliefs, reflections, or copyrighted/source material without realizing the skill retains artifacts on disk for future use.
The skill advertises searches of authoritative web sources but does not clearly disclose that user queries or local-topic-derived prompts may be sent over the network. This creates a privacy and data-handling risk because sensitive religious interests, personal questions, or content inferred from local materials could be exposed to external services without the user's informed consent.
The skill states it 'saves learned content to learned/ folder' and builds knowledge over time, but it does not clearly warn users that invoking the skill causes persistent local file writes. Silent persistence can surprise users, retain sensitive prompts or derived content on disk, and create downstream privacy or integrity risks if the learned folder is later synced, shared, or trusted as authoritative.
Suspicious Unicode normalization or mixed-script content
The flagged section contains explicit devotional dedication and lineage-centered religious framing, which can steer users toward a specific belief system without any indication of user consent, neutrality, or context gating. In an autonomous learning skill, this is risky because generated or surfaced content may present spiritual guidance as normative rather than optional, especially for vulnerable users seeking general informational help.
The document gives concrete psychedelic activation guidance, including substances, dose intensity cues, frequency, and progression framing, before presenting a consolidated safety warning. In a spiritual/self-development skill, this can normalize risky self-experimentation and may lead users with mental health vulnerabilities to attempt unsafe use despite scattered cautions later in the text.
The file includes a prescriptive religious practice script ('take refuge' and recite the eight precepts) presented as the default ceremony for users, without offering alternatives, opt-in framing, or sensitivity to user language, tradition, or beliefs. In a learning skill, this can create undue normative pressure and reduce user autonomy, especially for vulnerable or impressionable users seeking spiritual guidance.
The document explicitly fixes a specific religious-philosophical viewpoint and lineage in its framing, and later gives prescriptive practice guidance from that perspective. This can be a natural-language policy concern when a skill imposes one worldview or normative lens without user opt-in or alternatives, especially in advisory content.
This file explicitly fixes the skill's viewpoint to 'Middle Way (Prasangika Madhyamaka)', which can function as a forced perspective or locale-like policy constraint in natural-language behavior when presented as the default answer frame. Because the document does not offer alternative framing or user opt-in for other Buddhist traditions, it risks imposing a single viewpoint.