Back to skill

Security audit

粉丝投票共创剧情|AI-HIVE

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a disclosed AI-HIVE media-generation workflow, but its API clients can send saved bearer credentials to an arbitrary configured API base URL.

Review this skill before installing if you will use a real AI-HIVE API key. Use only the official AI-HIVE endpoint, avoid setting AI_HIVE_BASE_URL or --base-url from untrusted instructions, keep ~/.ai-hive/config.json private, and confirm prompts, uploads, batch size, routing, and cost before running generation commands.

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 (4)

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/imagegen.py:91
Finding

API Credential Disclosure Through an Unrestricted Base URL in the Image Generation Client

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/videogen.py:91
Finding

API Credential Disclosure Through an Unrestricted Base URL in the Video Generation Client

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Note
Location
scripts/imagegen.py:312
Finding

Unbounded Download of Server-Supplied Result URLs in the Image Generation Client

Content
View full analysis
0: pct = int(downloaded * 100 / total) bar = "█" * (pct // 5) + "░" * (20 - pct // 5) print(f"\r {bar} {pct}% ({downloaded // 1024}KB)", end="", flush=True) print() size_mb = downloaded / (1024 * 1024) print(f"[ok] {out_path} ({size_mb:.1f} MB)") except requests.exceptions.RequestException as e: print(f"\n[error] Download failed: {e}", file=sys.stderr) ``` ```python for i, item in enumerate(succeeded): result_url = item.get("resultUrl") if not result_url: continue ext = ".mp4" if "video" in task_type.lower() else ".png" filename = f"{task_type}_{task_id}_{i+1}{ext}" out_path = out_dir / filename download_file(result_url, out_path) for item in succeeded: last_frame = item.get("lastFrameUrl") if last_frame: out_path = out_dir / f"{task_type}_{task_id}_lastframe.png" download_file(last_frame, out_path) ``` ### Technical Analysis The client downloads `resultUrl` and `lastFrameUrl` values received from task responses without validating the URL scheme or destination host. It also follows the default redirect be ...[truncated 1655 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Note
Location
scripts/videogen.py:312
Finding

Unbounded Download of Server-Supplied Result URLs in the Video Generation Client

Content
View full analysis
0: pct = int(downloaded * 100 / total) bar = "█" * (pct // 5) + "░" * (20 - pct // 5) print(f"\r {bar} {pct}% ({downloaded // 1024}KB)", end="", flush=True) print() size_mb = downloaded / (1024 * 1024) print(f"[ok] {out_path} ({size_mb:.1f} MB)") except requests.exceptions.RequestException as e: print(f"\n[error] Download failed: {e}", file=sys.stderr) ``` ```python for i, item in enumerate(succeeded): result_url = item.get("resultUrl") if not result_url: continue ext = ".mp4" if "video" in task_type.lower() else ".png" filename = f"{task_type}_{task_id}_{i+1}{ext}" out_path = out_dir / filename download_file(result_url, out_path) for item in succeeded: last_frame = item.get("lastFrameUrl") if last_frame: out_path = out_dir / f"{task_type}_{task_id}_lastframe.png" download_file(last_frame, out_path) ``` ### Technical Analysis Video task result URLs are trusted without checking their scheme, hostname, redirect destination, response size, or media type. Streaming avoids loading the response into memory but ...[truncated 1353 chars]
Remediation
View remediation
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
Findings (24)

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

This mismatch includes undeclared user-info and wallet-balance queries, local config storage, direct upload/download, and broad AI generation functionality unrelated to the stated co-creation workflow. In context, that is more dangerous than a simple documentation bug because the skill also handles API keys and billable third-party operations, so users could unknowingly expose account metadata, store credentials insecurely, or trigger costly actions.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

This mismatch includes undeclared user-info and wallet-balance queries, local config storage, direct upload/download, and broad AI generation functionality unrelated to the stated co-creation workflow. In context, that is more dangerous than a simple documentation bug because the skill also handles API keys and billable third-party operations, so users could unknowingly expose account metadata, store credentials insecurely, or trigger costly actions.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

This mismatch includes undeclared user-info and wallet-balance queries, local config storage, direct upload/download, and broad AI generation functionality unrelated to the stated co-creation workflow. In context, that is more dangerous than a simple documentation bug because the skill also handles API keys and billable third-party operations, so users could unknowingly expose account metadata, store credentials insecurely, or trigger costly actions.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

This mismatch includes undeclared user-info and wallet-balance queries, local config storage, direct upload/download, and broad AI generation functionality unrelated to the stated co-creation workflow. In context, that is more dangerous than a simple documentation bug because the skill also handles API keys and billable third-party operations, so users could unknowingly expose account metadata, store credentials insecurely, or trigger costly actions.

Content

No source excerpt is available for this finding.

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 81)May include surrounding context.

md
python3 scripts/imagegen.py init --skill-name fan-vote-story-co-creation-ai-hive

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 83)May include surrounding context.

md
python3 scripts/imagegen.py init --skill-name fan-vote-story-co-creation-ai-hive

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
95% confidence
Finding

The skill advertises and demonstrates capabilities that involve environment variables, local file access, shell execution, and network/API usage, but it does not declare any explicit tool scope or permission boundaries. This creates an over-privileged integration risk: a host agent may grant broader capabilities than intended, making accidental file access, command execution, or outbound data transfer easier if the skill is invoked in sensitive contexts.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The manifest description and title are entirely framed around Chinese trigger phrases and Chinese-language use, and the examples throughout the document assume Chinese prompts and outputs. There is no explicit opt-in, multilingual option, or statement that the skill is intentionally limited to a Chinese-only regional compliance context.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The default prompt hardcodes Chinese output ('production-ready Chinese workflow') without checking user preference. While not directly enabling code execution, this can override user intent, reduce transparency, and make safety-relevant instructions less understandable to some users or downstream reviewers, which is more concerning in a skill that emits runnable commands and operational deliverables.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The manifest sets allow_implicit_invocation: true, which permits the skill to be invoked automatically without a tightly scoped trigger boundary in the file itself. Because this skill can generate production-ready workflows and runnable AI-HIVE commands, overly broad implicit activation increases the chance of unintended invocation on adjacent user requests, causing unreviewed content generation or API-oriented actions in a higher-risk marketing/content-production context.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This code embeds Chinese-only natural-language strings in the module docstring, argument parser description, skill name, and search aliases, with no indication that the user can choose another language or that the skill is intentionally limited to a Chinese-speaking context. That creates a locale policy concern because the skill appears to mandate a specific language by default rather than offering user choice.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/edit_video.py (reported line 20)May include surrounding context.

python
def run(args: list[str]) -> None:
    print(" ".join(args))
    subprocess.run(args, check=True)


def main() -> None:

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The file is packaged as a narrowly scoped fan-vote story image skill, but it actually embeds a generic AI-Hive client with broader capabilities. This scope mismatch increases the attack surface and can enable unintended operations through a skill that users and reviewers may reasonably trust as single-purpose.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

This code file contains natural-language descriptions and user-facing CLI messages that assume Chinese as the only language of interaction. Under the stated policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale restriction is explicitly justified, which is not present here.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

Exposing a user-info and wallet-balance command gives the skill access to account metadata unrelated to image generation. In a delegated-agent setting, this can leak sensitive account information or enable reconnaissance about available funds without a clear business need.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The embedded text-chat capability is outside the declared image-focused fan-vote workflow and materially expands what the skill can do. In practice, this can be abused to relay arbitrary prompts and data to external models, bypassing the narrower controls and expectations attached to the packaged skill.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The file presents itself as a fan-vote story co-creation video skill, but it also includes broad OpenAPI functionality for chat, image generation, model enumeration, user-info, uploads, and generic task handling. This scope expansion violates least privilege and increases the chance that callers can use the skill for unrelated actions, data exposure, or policy bypass beyond the declared purpose.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill exposes a user-info endpoint that can reveal account metadata and wallet balance, which is unrelated to fan-vote video generation. Unnecessary access to account information expands the blast radius if the skill is misused or invoked in an unintended context.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The generic text chat command is outside the declared specialized video co-creation purpose and enables arbitrary model interactions through the same trusted skill package. That broader capability can be abused to send unrelated prompts or process uploaded media in ways users and reviewers would not expect from this skill.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The image-generation pathway is broader than the skill's declared fan-vote story video function and allows creation of standalone image outputs through the same package. This creates an unnecessary capability surface that may be used for unrelated content generation, complicating review and weakening trust in the skill boundary.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill uploads local media files to remote storage and API services, but the interface does not give an explicit warning at the point of use that selected files will leave the local machine. In a creative workflow, users may upload sensitive or proprietary assets, so lack of clear disclosure can cause unintended data exfiltration to third-party infrastructure.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The top-level docstring says this file is a '通用 AI 能力调用工具' that wraps text chat, image generation, video generation, model querying, media upload, and task polling. However, the generated skill configuration at the bottom packages the file as the specific fan-vote-story-co-creation-ai-hive image skill, so the documentation actively describes a broader intent than the delivered skill claims.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
96% confidence
Finding

The top-level natural-language description and many user-facing strings are Chinese-only, which effectively imposes a locale/language requirement without offering user opt-in or alternatives. Under the stated policy, forcing a specific language without choice is a natural-language policy concern unless clearly justified as region-specific.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

The script writes downloaded outputs to user directories and also creates the output directory if needed. While runtime progress messages show the filename, there is no broader disclosure in the skill description or command help that running generation will create or overwrite local files under the configured output directory.

Content

No source excerpt is available for this finding.