Back to skill

Security audit

San Francisco

Security checks for vulnerabilities and agentic risk

Overview

This is a static San Francisco advice guide with no executable behavior, though users should verify time-sensitive safety, cost, and travel details.

Installing this skill appears low risk from a security standpoint. Use it as general San Francisco background, and verify current prices, transit status, safety conditions, emergency guidance, and legal or government requirements before acting on them.

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
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The phrase 'User asks about San Francisco for any purpose' covers an extremely wide range of everyday queries and does not clearly define the boundaries of when this skill should or should not activate. The description also lacks explicit exclusions or negative examples, increasing the risk of unintended invocation for general conversation about San Francisco.

Content

No source excerpt is available for this finding.

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

Medium
Category
YARA Match
Confidence
65% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · resident.md (reported line 67)May include surrounding context.

md
e, owned by Zillow
5. **Facebook Groups** — Roommates and subleases
6. **Walking neighborhoods** — Many listings only have "For Rent" signs

### Application Documents (Have Ready)
- Proof of income: **3x monthly rent** (offer letter, pay stubs)
- Credit report or score
- Bank statements (3 months)
- References (previous landlords)
- Photo ID
- **Be ready to apply same day** — Competition is fierce

### Scam Warning Signs
- Below-market rent for area
- Request for wire transfer
- Can't view in person
- Listing "too good" for location
- Pressure to send money fast

## Cost of Living (Monthly)

### Single Person: $4,500-5,500/month

| Category | Monthly Cost |
|----------|-------------|
| Rent (1BR downtown) | $3,000-3,500 |
| Rent (1BR outer neighborhoods) | $2,400-2,800 |
| Utilities (PG&E) | $150-255 |
| Internet | $65-80 |
| Groceries | $400-800 |
| Transport (Clipper pass) | $81-98 |
| Phone | $50-70 |
| **Total (without extras)** | **$4,200-5,300** |

### Additional Costs
- He

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

The heading "Richmond District ("Better Chinatown")" includes a value-laden comparative label that could be interpreted as prescribing a preferred cultural framing rather than neutrally describing the area. This is a natural-language quality/policy concern because it embeds an unsupported and potentially insensitive characterization in the skill content.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

This guide gives location-specific safety and emergency-related advice in a highly authoritative tone, including when to avoid areas, how to respond to people in crisis, and what to do during earthquakes, without warning users that conditions can change rapidly and that official real-time guidance should take precedence. If information is outdated, overgeneralized, or interpreted as definitive, users could make unsafe decisions based on it.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.