Sydney

• Outcome: Pass
• Latest audit: Sep 10, 2026 · 8:47 PM UTC
• Version: 1.0.0

Overview

This is a static Sydney guide with no hidden execution, persistence, credential access, or install-time behavior; users should still verify time-sensitive and housing-related advice.

Install only if you want a static Sydney reference guide. Treat visa, pricing, transport, legal, healthcare, and market figures as dated February 2026 and verify with official sources before making decisions. For neighborhood advice, rely on concrete factors like commute, budget, amenities, schools, and safety rather than demographic labels in the text.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• YARA Signatures: Malware Match, Webshell Match, Cryptominer Match
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (21)

Anti-Refusal Statement

High

• Category: Anti-Refusal

Content

**Unmarked cars with cameras inside.** They park on roadsides, in medians, on overpasses.

- **No warning signs** (NSW removed mandatory signage)
- Can be anywhere, anytime
- Use Waze app for community reports (not guaranteed)

• Confidence: 80% confidence
• Finding: 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.

Vague Triggers

Medium

• Confidence: 95% confidence
• Finding: This markdown file's invocation guidance says the skill should be used when a user asks about Sydney 'for any purpose,' which is broad enough to overlap with many ordinary conversations and unrelated intents. It does not provide narrower trigger phrases, constraints, or negative examples to distinguish when the skill should or should not activate.

YARA rule 'network_reconnaissance': Network reconnaissance and scanning patterns [hacktools]

Medium

• Category: YARA Match

Content

se** — easily 40–50% of income in the city
2. **Electricity is genuinely expensive** — budget for summer A/C spikes
3. **Alcohol is taxed heavily** — 2–3x European/US prices
4. **Medicare is excellent** — if you qualify (citizens/PR only)
5. **Public transport has weekly caps** — use it, don't drive to CBD
6. **CBD parking is brutal** — $50–80/day
7. **Competition for rentals is fierce** — 50+ applications per property in good areas
8. **Sydney is expensive but not NYC/London expensive** — especially healthcare
9. **Aldi saves real money** — 20–40% cheaper on groceries
10. **Free beaches, parks, hiking** — lifestyle offsets some costs

---

*Last updated: February 2026*
*Sources: Numbeo, Transport NSW, realestate.com.au, local market research*

• Confidence: 65% confidence
• Finding: YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Natural-Language Policy Violations

Medium

• Confidence: 92% confidence
• Finding: The content uses broad community-based framing such as describing areas as 'Chinese-focused' or otherwise reducing suburbs to demographic identity. In a neighborhood selection guide, this can steer users using stereotypes rather than neutral, relevant criteria and may alienate protected groups or encourage biased housing decisions.

Natural-Language Policy Violations

Medium

• Confidence: 96% confidence
• Finding: Labeling a suburb primarily by religion, such as 'Muslim community,' introduces protected-attribute characterization into housing-style guidance. In this context, that can facilitate discriminatory filtering by users seeking to avoid or target areas based on residents' religion rather than legitimate lifestyle factors.

Natural-Language Policy Violations

Medium

• Confidence: 94% confidence
• Finding: This content explicitly frames suburbs through racial and ethnic composition in a way that can steer housing decisions based on protected characteristics. In a neighborhood recommendation skill, that context makes the issue more dangerous because users may rely on it for relocation choices, creating fair-housing style steering risks and discriminatory outcomes.

Natural-Language Policy Violations

Medium

• Confidence: 96% confidence
• Finding: These lines appear to promote a suburb using concentrated ethnic identity such as a 'Little Hong Kong' framing and 'strong community for Chinese speakers,' which can encourage users to select housing based on race, ethnicity, or national origin. Because the skill is specifically designed to guide visitors, residents, and workers on where to live, this is not merely descriptive travel content; it can function as discriminatory housing steering.

Natural-Language Policy Violations

Medium

• Confidence: 97% confidence
• Finding: Describing a suburb as 'very white, not diverse' uses race-based exclusionary wording that can influence where users feel welcome or unwelcome. In a relocation and neighborhood-selection guide, this materially increases the risk of discriminatory steering by encoding racial composition as a decision factor.

Natural-Language Policy Violations

Low

• Confidence: 81% confidence
• Finding: The phrase "Australia's equivalent of Michelin stars" and the surrounding explanation assume an Australian restaurant-rating framework without indicating that the content is specific to Australia/Sydney dining conventions. This is a minor natural-language locale policy concern because it presents a region-specific standard as the default reference point without explicit user choice or opt-in.

Missing User Warnings

Low

• Confidence: 90% confidence
• Finding: The guide provides extensive alcohol venue recommendations, BYO advice, and late-night drinking suggestions while offering little practical harm-reduction guidance beyond RSA rules. In a city-navigation skill, this can normalize heavy or late-night drinking without reminding users about intoxication risks, drink spiking, transport planning, or drink-driving laws, which creates a genuine but low-severity safety issue.

Natural-Language Policy Violations

Low

• Confidence: 76% confidence
• Finding: This markdown file is entirely written as end-user guidance in English and does not indicate that users may request another language or translated support. For a public-facing healthcare skill, that can be a locale/language policy concern if organizational policy requires user choice rather than a forced default language.

Natural-Language Policy Violations

Low

• Confidence: 89% confidence
• Finding: The statement "Australians say 'bushwalking,' not hiking" prescriptively directs the user toward a specific local language convention. This is a locale-specific language preference presented as a rule, without offering user choice or framing it as optional local context.

Natural-Language Policy Violations

Low

• Confidence: 83% confidence
• Finding: The line instructs that "How are you" should be treated as a greeting with a specific expected response, which imposes a locale-specific communication norm. It does not present this as optional or user-selectable, so it can be read as forcing a particular language behavior.

Natural-Language Policy Violations

Low

• Confidence: 86% confidence
• Finding: The document is explicitly framed as a guide for Sydney neighborhoods, which fixes the locale to a specific city and region. Under the stated policy, locale constraints should either be user-selectable or clearly justified; this file presents the locale as mandatory without any opt-in language.

Natural-Language Policy Violations

Low

• Confidence: 77% confidence
• Finding: Line L291 describes the area as suitable for 'International students (familiar culture, food)', which implicitly centers a particular cultural/linguistic background without offering user choice or clarifying scope. In a skill context, this can amount to a natural-language locale/culture assumption rather than neutral guidance.

Natural-Language Policy Violations

Low

• Confidence: 81% confidence
• Finding: Advising users to avoid 'very conservative suburbs' is a value-laden generalization that can encode political or social bias into location guidance. While less severe than protected-attribute labeling, it still encourages exclusionary reasoning and weakens the neutrality expected in relocation advice.

Natural-Language Policy Violations

Low

• Confidence: 81% confidence
• Finding: Line L079 states 'Competent English (IELTS 6.0+ each band minimum)' as a requirement for Subclass 189, which reads as a mandatory specific-language requirement. Later at L306-L308, the document acknowledges exceptions, so this earlier phrasing is overly absolute and could misstate language policy requirements without user-specific qualification.

Missing User Warnings

Low

• Confidence: 87% confidence
• Finding: The skill recommends specific swimming spots at Bondi without any basic warning about surf conditions, flags, rip currents, or following lifeguard guidance. In a travel skill, users may treat these as actionable safety advice, so omission could contribute to unsafe behavior in hazardous ocean conditions.

Missing User Warnings

Low

• Confidence: 82% confidence
• Finding: The Watsons Bay section recommends The Gap and the South Head Heritage Trail but omits any caution about staying behind barriers, supervising children, and avoiding cliff edges in high wind or rough weather. Because this is destination guidance encouraging visitation to elevated coastal lookouts, the lack of a basic warning increases risk of accidental injury.

Missing User Warnings

Low

• Confidence: 78% confidence
• Finding: The Hunter Valley section recommends a wine-tour day built around multiple tastings and includes a self-drive option, but only briefly says to designate a driver. Because this markdown file describes behavior that can affect user safety, it should more clearly warn about alcohol consumption, transport planning, and avoiding driving after tastings.

Natural-Language Policy Violations

Low

• Confidence: 83% confidence
• Finding: The description states the advice is provided "from a local's perspective," which frames the content around a single locale-specific viewpoint without offering alternatives or clarifying that this is intentionally region-specific guidance. Under the language/locale policy rule, this can be read as imposing a specific locale perspective rather than offering user choice.

Static analysis

No suspicious patterns detected.