Back to skill

Security audit

Aries Holiday Weekday

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a coherent China holiday checker, but it exposes an arbitrary URL fetch path and publishes credential-like gov.cn signing material that users should review before installing.

Install only if you are comfortable with a helper script that can make outbound web requests and write skill-local cache files. The maintainer should restrict notice fetching to official gov.cn HTTPS hosts, remove or replace captured signing material with an authorized configuration path, and keep the cache behavior clearly documented.

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/query_holiday.py:236
Finding

Unrestricted URL Fetching Enables Server-Side Request Forgery

Content
View full analysis
str: """Fetch and extract the main body text from a gov.cn notice page.""" response = requests.get( url, headers={ "User-Agent": "Mozilla/5.0", "Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8", }, timeout=20, ) response.raise_for_status() response.encoding = response.apparent_encoding or response.encoding or "utf-8" html = response.text text = _extract_div_text( html, lambda attrs: "trs_editor_view" in (attrs.get("class") or "").split(), ) if not text: text = _extract_div_text(html, lambda attrs: attrs.get("id") == "UCAP-CONTENT") if not text: raise SystemExit(f"Could not extract notice body from {url}") return text ``` The URL is also exposed directly through the CLI: ```python body_parser = sub.add_parser("body", help="Fetch a notice body") body_parser.add_argument("url") ``` ```python if args.command == "body": print_body(args.url) return ``` ```python def print_body(url: str) -> None: print(fetch_notice_body(url)) ``` ### Technical Analysis The `body` command accepts an arbitrary URL and passes it directly to `requests.get()`. The implementation does not validate: - The URL scheme - The destination hostname - The resolved IP address - The destination port - Redirect destinations - Whether the target belongs to the intended `gov.cn` domain Although the documented purpose is to retrieve official government notices, this trust boundary is not enforced in code. The `requests` library also follows redirects by default, so even an initially acce ...[truncated 1761 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/query_holiday.py:24
Finding

Hard-Coded API Key and Signing Material Exposed in Source and Documentation

Content
View full analysis
str: return md5_hex(f"{ATHENA_KEY}{ATHENA_SECRET}{timestamp_ms}TRS") ``` ```python return { "User-Agent": "Mozilla/5.0", "Content-Type": "application/json;charset=UTF-8", "Accept": "application/json, text/javascript, */*; q=0.01", "Origin": BASE_URL, "Referer": referer, "athenaAppKey": ATHENA_APP_KEY, "athenaAppName": ATHENA_APP_NAME, "T-KEY": ATHENA_KEY, "T-SEC": ATHENA_SECRET, "T-T": str(timestamp_ms), "T-SIGN": build_sign(timestamp_ms), } ``` The documentation also publishes the signing values: ```markdown - `athenaAppKey`: captured encoded key from the browser session - `athenaAppName`: `%E5%9B%BD%E7%BD%91%E6%90%9C%E7%B4%A2` - `T-KEY`: `irs-c-box` - `T-SEC`: `99b2eeecb4f10339` ``` ### Technical Analysis The project embeds an application key and signing material directly in executable source and repeats part of that material in documentation. Anyone with access to the package can recover the values and reproduce the signature algorithm. MD5 is used here as a deterministic digest rather than a secret-preserving credential mechanism. Because the signing secret and exact input ...[truncated 1791 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (9)

Direct Prompt Extraction

High
Category
System Prompt Leakage
Confidence
85% confidence
Finding

Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.

Content

Scanner excerpt · README.md (reported line 17)May include surrounding context.

md
- [`SKILL.md`](SKILL.md)
- [`scripts/query_holiday.py`](scripts/query_holiday.py)

## Output rules

- Do not expose internal cache or refresh state in user-facing answers
- If a progress line is needed, use only:

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
95% confidence
Finding

The skill directs the agent to read local cache files, write refreshed year data, and fetch data from gov.cn, but it does not declare any explicit tool scope or allowed-tools boundary. That creates an authorization gap where the runtime may permit broader file or network actions than intended, increasing the blast radius if the skill is modified, misinterpreted, or composed with other agent behaviors.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The workflow requires the agent to say 正在获取{年份}年最新节假日安排, and later output-selection prompts are also written as fixed Chinese instructions. This imposes a specific language on user-visible behavior without documenting a user choice or opt-in, which matches the language/locale policy violation criteria.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The reference embeds multiple hard-coded credential-like values, including an app key, signing key name, and a secret used to generate request signatures. Even if these were originally captured from a browser session, publishing them in skill documentation encourages reuse of potentially unauthorized access material and can enable abuse of the upstream gov.cn search API, quota theft, or future breakage if the values are revoked.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The body subcommand accepts an arbitrary URL and passes it directly to requests.get, allowing the skill to fetch and return content from any location rather than only official holiday notices. In an agent setting, this expands the skill into a general network retrieval primitive, which can be abused for unintended data access, internal endpoint probing, or policy bypass if user input is not tightly constrained.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
87% confidence
Finding

This markdown file states that the skill sends requests to the official government website when refreshing yearly holiday data, but it presents this as a usage suggestion rather than a clear user warning about network activity. Under the markdown-file criteria, behavior that could affect privacy or system integrity should be explicitly warned about, and the current wording does not clearly disclose that using the skill may initiate outbound network access.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

This JSON file contains user-facing natural-language fields entirely in Chinese, including the notice title and source URL context, with no accompanying indication that the skill is China-specific or that users can opt into this locale. Under the policy, forcing a specific language without user choice can be a locale-policy concern.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The code exposes generic gov.cn search and page-retrieval helpers that go beyond a narrowly scoped holiday weekday checker. While limited to the gov.cn ecosystem rather than arbitrary internet access, this still broadens the skill's effective capability and can let users retrieve unrelated government content not described in the manifest.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

This code writes holiday data to a local JSON cache file, which is a file-modifying operation. Although the function name implies caching, there is no confirmation prompt, log/print notice, or inline comment/docstring here informing the user that running refresh behavior will create or overwrite files under the cache directory.

Content

No source excerpt is available for this finding.