Back to skill

Security audit

scholar-search-zh

Security checks for vulnerabilities and agentic risk

Overview

This looks like a legitimate AISA search skill, but it exposes broader web, URL extraction, deep research, and synthesis features than its academic-search description clearly tells users.

Install only if you are comfortable sending search queries, prompts, URLs, aggregated search results, and your AISA API token to AISA's API. Treat it as a broader multi-search and research client, not just an academic-paper lookup skill, and avoid using it with private/internal URLs or confidential research topics unless that disclosure is acceptable.

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
Findings (1)

other

Note
Location
scripts/search_client.py:125
Finding

Undeclared Network Capabilities Exceed the Skill's Academic Search Scope

Content
View full analysis
None: api_key = get_api_key() body: dict[str, Any] = { "query": args.query, "max_results": args.count, "search_depth": args.depth, } if args.topic: body["topic"] = args.topic if args.time_range: body["time_range"] = args.time_range if args.include_answer: body["include_answer"] = True data = aisa_post(api_key, "/tavily/search", body) print_results(data, "Tavily Search") def cmd_extract(args: argparse.Namespace) -> None: api_key = get_api_key() urls = [u.strip() for u in args.urls.split(",")] data = aisa_post(api_key, "/tavily/extract", {"urls": urls}) if "results" in data and isinstance(data["results"], list): for r in data["results"]: print(f"\n{'='*60}") print(f" URL: {r.get('url', 'Unknown')}") print(f"{'='*60}") content = r.get("raw_content", "") print(content[:3000] if content else "(no content)") else: print(json.dumps(data, indent=2)) def cmd_sonar(args: argparse.Namespace) -> None: api_key = get_api_key() endpoint_map = { "sonar": "/sonar", "sonar-pro": "/sonar-pro", "sonar-reasoning-pro": "/sonar-reasoning-pro", "sonar-deep-research": "/sonar-deep-research", ...[truncated 7759 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (11)

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The skill is presented as a narrowly scoped academic search tool, but the detected behavior includes general web search, arbitrary URL extraction, external research-answering endpoints, and multi-source aggregation beyond the declared purpose. This mismatch is dangerous because users and orchestrators may grant trust or invoke the skill under the assumption of limited scholarly retrieval, while it actually has much broader data-fetching and network reach.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding

The skill declares access to an API key and performs networked operations, but it does not define any explicit tool scope such as allowed tools or permissions. That makes its operational boundary ambiguous and increases the risk of overbroad execution or unintended access paths when invoked by an agent runtime.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The skill name, title, and all user-facing instructions are presented only in Chinese, which can constitute a language/locale policy violation when no user opt-in or alternative language option is provided. The file does not state that the skill is region-specific or otherwise justify the language restriction.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest sets the skill language to "zh", which indicates a fixed locale. In this file there is no indication that the user can opt into another language or that the locale restriction is justified as region-specific, which matches the language/locale policy violation criteria.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The file advertises and implements capabilities well beyond the declared scholar-search scope, including general web search, extraction, Sonar research, and multi-source synthesis. Scope expansion is dangerous because it increases data exposure and grants operators undeclared capabilities that users and reviewers would not reasonably expect from an academic-search skill.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/search_client.py (reported line 28)May include surrounding context.

python
from concurrent.futures import ThreadPoolExecutor, as_completed
from typing import Any

AISA_BASE = "https://api.aisa.one/apis/v1"


def get_api_key() -> str:

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

User queries and URLs are sent directly to external endpoints without any explicit warning, consent, or privacy notice in the client flow. This is risky because search terms may contain confidential research topics, personal data, or internal URLs, and the skill context suggests academic lookup rather than broad third-party disclosure.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

Arbitrary URL extraction lets the skill transmit user-supplied URLs to an external service and retrieve large bodies of remote content unrelated to scholarly lookup. In a scholar-search context, this creates an unnecessary exfiltration and misuse surface, especially if users provide sensitive, private, or internal links under the assumption they are only doing paper search.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The Sonar commands expose a general deep-research/chat capability rather than a narrow scholarly search function. This broadens the skill into an undeclared LLM research interface, increasing the chance that user prompts and sensitive research topics are sent to services outside the expected academic-search workflow.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The verity workflow aggregates multiple external searches, computes confidence scores, and sends combined results for AI synthesis, which materially exceeds a simple scholar-search role. This increases the volume and sensitivity of transmitted data and may present synthesized conclusions as trustworthy without clear provenance or user expectation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The explain request hard-codes "language": "en", which imposes an English-only output policy regardless of user preference. The file does not provide an argument, configuration, or opt-in mechanism to select another language, so this is a natural-language locale policy violation.

Content

No source excerpt is available for this finding.