Back to skill

Security audit

Daily Briefing

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a daily briefing tool, but it needs review because one input can trigger local command execution and the documentation includes a scheduled email example to a fixed third-party address.

Install only if you are comfortable running local Node scripts that contact weather/news sites. Do not pass untrusted city names until the shell command is replaced with safe HTTP fetching, and do not enable the README email cron example unless the recipient is changed to an address you control and recurring delivery is explicitly approved. The Beijing traffic rules are also hard-coded for 2025-12-29 through 2026-03-29, so they are stale as of 2026-09-16.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/data-collector.mjs:39
Finding

OS Command Injection Through Unsanitized City Input

Content
View full analysis
/dev/null || echo '{}'`; const output = execSync(cmd, { encoding: 'utf8', timeout: 10000 }); ``` The `city` value originates from the exported `getWeather(city)` API and can also be supplied through higher-level exported functions such as `generateBriefing({ city })`. ### Technical Analysis The application constructs a shell command by directly interpolating the `city` argument into a single-quoted URL and then passes the resulting string to `execSync()`. No allowlist validation, shell escaping, or safe argument separation is applied. Because `execSync()` invokes a shell for a string command, an attacker can supply a city value containing a single quote to terminate the quoted URL, followed by shell metacharacters and an arbitrary command. The trailing portion of the original command can then be neutralized with a shell comment or otherwise incorporated into the injected command. This vulnerability is reachable through the project's public JavaScript API even though the current command-line interface does not expose a city argument directly. ### Attack Path 1. An application imports `getWeather()`, `generateBriefing()`, or another API that accepts the `city` option. 2. The application passes a city value influenced by an untrusted user. 3. The malicious value closes the single-quoted URL and appends an operating-system command. 4. `getWeather()` incorporates the value into the shell command without validation. 5. `execSync()` executes the resulting command under the identity and privileges of the Node.js process. 6. The injected command can access any files, environment variables, credentials, or network resources available to that process. Conceptually, an input shaped like the following demonstrates the injectio ...[truncated 860 chars]
Remediation
View remediation

other

Warning
Location
README.md:178
Finding

Hard-Coded Third-Party Email Recipient in Cron Integration Instructions

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (27)

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 23)May include surrounding context.

md
node scripts/generate-briefing.mjs

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 26)May include surrounding context.

md
node scripts/generate-briefing.mjs

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 29)May include surrounding context.

md
node scripts/generate-briefing.mjs

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 32)May include surrounding context.

md
node scripts/generate-briefing.mjs

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 35)May include surrounding context.

md
node scripts/generate-briefing.mjs

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 58)May include surrounding context.

md
node scripts/generate-briefing.mjs

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 142)May include surrounding context.

md
node scripts/generate-briefing.mjs

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 140)May include surrounding context.

md
- `scripts/data-collector.mjs` - 天气、限行数据获取

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 141)May include surrounding context.

md
- `scripts/news-search.mjs` - 新闻搜索与分类整理模块

Known Vulnerable Dependency: undici==7.24.4 — 12 advisory(ies): CVE-2026-6733 (undici vulnerable to HTTP response queue poisoning via keep-alive socket reuse); CVE-2026-13697 (undici vulnerable to cross-user information disclosure and parse-time crash via ); CVE-2026-16728 (undici vulnerable to downstream response desynchronization via retry interceptor) +9 more

High
Category
Supply Chain
Confidence
94% confidence
Finding

The lockfile pins transitive dependency undici to version 7.24.4, and the supplied advisory list indicates multiple known HIGH-severity vulnerabilities affecting that exact version. Because this skill generates briefings by collecting external data, it likely performs outbound HTTP requests; a vulnerable HTTP client in that path can enable response desynchronization, queue poisoning, information disclosure, or denial of service when interacting with malicious or compromised upstream endpoints.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

SQP-3 applies to all file types and covers language/locale policy violations. The README and sample outputs are exclusively Chinese, and the cron email example explicitly formats dates with toLocaleDateString('zh-CN'), but there is no indication that the user can choose another language/locale or that the skill is intentionally limited to a China-specific audience.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The README’s Cron example expands the skill from generating a local briefing into sending email via a separate skill, introducing outbound data exfiltration capability that is not part of the stated daily-briefing purpose. In an agent ecosystem, users may copy this configuration directly, causing generated content to be automatically transmitted off-host without a clear trust boundary or consent warning.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
95% confidence
Finding

The skill documents direct shell execution (node ...mjs) and a cron payload that instructs an agent to run local scripts, but it does not declare any explicit tool scope such as allowed shell access. This creates a permissions/documentation mismatch: an operator or platform may treat the skill as lower risk than it is, while the skill actually requires command execution capabilities against local files under the user's home directory.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The file presents all user-facing instructions and operational details in Chinese, including setup, usage, and cron message content. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale constraint is explicitly justified, which is not stated here.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The weather lookup builds a shell command with the untrusted city parameter interpolated directly into curl -s 'wttr.in/${city}?format=j1', then executes it with execSync. Because the value is placed inside single quotes, an attacker can inject a quote and shell metacharacters to achieve command injection, which is far more dangerous than the skill’s simple data-collection purpose would justify.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The code explicitly formats dates with the 'zh-CN' locale and throughout the file returns Chinese-only user-facing strings, including comments indicating the tool is for a Chinese morning briefing. This imposes a specific language/locale on all users without offering a choice or documenting a justified regional-only constraint.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

This code formats dates using the zh-CN locale and the surrounding output strings are also in Chinese, which effectively forces a specific language/locale. The file does not offer a user opt-in or configurable locale, and there is no documented justification here for restricting output to Chinese.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The generation timestamp is explicitly formatted with zh-CN, continuing the pattern of enforcing one locale across user-visible output. Because no alternative locale path or opt-in is provided, this is a natural-language policy concern rather than a technical bug.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The file description, category labels, status messages, and returned content are written for Chinese-language news collection and output, but there is no indication that the user can opt into this locale or choose another language. This creates a natural-language policy issue because the skill effectively forces a specific language experience.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The manifest describes a daily briefing skill that collects weather, traffic, and news data; network retrieval is expected, but spawning a shell with execSync is not an obvious or necessary capability for that purpose. Using child_process introduces command-execution capability beyond simple data collection, which is context-inappropriate for a briefing generator.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

This code invokes curl through execSync to make outbound requests, which is a safety-relevant operation for a code file because it executes a shell command and transmits data to external sites. While there are console logs about fetching, they do not clearly warn a user that the script will run subprocesses and contact multiple external news sources.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
87% confidence
Finding

The documentation promotes automatic fetching from external weather/news sources and shows an email notification workflow, but it does not clearly disclose outbound network access, what data leaves the environment, or that generated briefings may be transmitted to third parties. This can mislead operators in restricted or privacy-sensitive deployments and increases the risk of unintended data disclosure through routine scheduled execution.

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

This manifest file describes the skill only as generating 'daily morning briefings' but does not define any specific invocation phrase, scope, or exclusion conditions. In a manifest context, that broad natural-language description can be too vague to distinguish when the skill should activate versus when a general request about weather or news should not invoke it.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
84% confidence
Finding

Using a caret range for a dependency allows newly published minor or patch releases to be installed automatically, which can introduce vulnerable or malicious upstream code without review. In a skill that gathers external content for automated briefings, supply-chain compromise could affect every execution and any downstream data processing.

Content

Scanner excerpt · package.json (reported line 11)May include surrounding context.

json
"test": "node scripts/test-all.mjs"
  },
  "dependencies": {
    "cheerio": "^1.2.0"
  },
  "keywords": ["daily", "briefing", "weather", "news", "beijing"],
  "author": "Kimi Claw",

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
75% confidence
Finding

The keyword list includes 'beijing', which suggests the skill may be tailored to a specific locale, but the manifest provides no indication that users can choose or opt into that locale. If the skill behavior is locale-specific, this should be explicitly documented and presented as a user choice or justified region-specific scope.

Content

No source excerpt is available for this finding.