Security audit

Briefing Diario

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Portuguese daily briefing that fetches weather, exchange-rate, holiday, and time data without installing code, persisting state, or accessing credentials.

Install this only if you want a Brazil-focused Portuguese daily briefing. Be aware it may contact public weather, exchange-rate, and holiday APIs, and consider invoking it with explicit phrases to avoid accidental runs from generic words like "dia".

A.I.G

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

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Vague Triggers

Medium

Confidence: 94% confidence
Finding: A skill that auto-invokes on broad phrases like 'briefing diário' or 'briefing' can be triggered unintentionally during normal conversation, causing unrequested tool use and data fetching. In this skill's context, accidental invocation can also initiate external requests and location-based processing without clear user intent.

Natural-Language Policy Violations

Medium

Confidence: 96% confidence
Finding: The instructions require translating the month to Portuguese, and later sections also require Brazilian Portuguese weather and moon-phase text. This enforces a specific language/locale without opt-in or explanation that the skill is intentionally limited to Portuguese-speaking users.

External Transmission

Medium

Category: Data Exfiltration
Content: **Clima e Astronomia:** Use o endpoint da Open-Meteo: ``` https://api.open-meteo.com/v1/forecast?latitude=-19.9208&longitude=-43.9378&daily=uv_index_max,sunset,sunrise&hourly=precipitation_probability&current=temperature_2m,relative_humidity_2m,apparent_temperature,weather_code,wind_speed_10m&timezone=auto ``` Extraia e mapeie:
Confidence: 84% confidence
Finding: The skill directs the agent to call third-party APIs for weather, exchange rates, and holidays, which creates external data transmission and dependency risks. Although the weather URL shown uses fixed coordinates, the skill also instructs the agent to identify the user's city, making the overall design potentially privacy-sensitive if real user location is later inserted into outbound requests.

Natural-Language Policy Violations

Medium

Confidence: 95% confidence
Finding: This line instructs the agent to translate conditions to "Português do Brasil" as a mandatory behavior. Under the policy, forcing a specific language or locale is a violation unless the user is given a choice or the regional limitation is clearly justified.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The additional triggers 'dia', 'resumo do dia', and 'briefing' are vague everyday phrases that are likely to appear in unrelated conversation, increasing the chance of accidental activation. Because this skill performs multiple data lookups and derives user-local context, unintended invocation has privacy and safety implications beyond mere formatting errors.

Static analysis

No suspicious patterns detected.