ClawHub
Back to skill

Security audit

infinite-pokemon-npc

Security checks for vulnerabilities and agentic risk

Overview

This skill is mainly a scoped game NPC generator, but its optional setup path can clone and run an unpinned external game repository.

Use the NPC-generation behavior only with host-supplied game context. Before asking the skill to set up or launch the full game, review the upstream repository and README, prefer a known commit or release, and run dependency installation/build steps in an isolated environment without unnecessary credentials or sensitive files exposed.

A.I.G

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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 (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:17
Finding
Unpinned External Repository Setup Workflow## Vulnerability Details **File Location**: `SKILL.md`, lines 17–22 **Vulnerability Type**: Supply-chain risk caused by retrieving and executing mutable external content **Risk Level**: Medium ### Vulnerable Code Snippet ```markdown Installing this skill alone does not install Infinite Pokémon. If the game checkout or any required project file is missing, first try to clone the complete [GitHub repository](https://github.com/Shellishack/infinite-pokemon) into a new local directory before starting play: ```sh git clone https://github.com/Shellishack/infinite-pokemon.git cd infinite-pokemon ``` Read the cloned repository's `README.md` (or `README.zh-CN.md`) and follow its getting-started instructions to install dependencies, build, launch, and connect Codex. ``` ### Technical Analysis The setup workflow clones the repository's mutable default branch without pinning a reviewed commit hash, signed tag, or release artifact. It then instructs the agent to follow setup instructions retrieved from that repository, including dependency installation, building, and launching software. Consequently, the effective commands and dependencies executed during setup can change after this skill bundle has been reviewed. Compromise of the upstream repository, its maintainer account, release process, or a dependency referenced by its setup instructions could introduce attacker-controlled code without requiring any modification to this audited package. This is an external setup path rather than evidence that the currently audited files contain malicious executable code. ### Attack Path 1. An attacker compromises the upstream repository, its default branch, or an account authorized to modify it. 2. The attacker changes the repository's setup documentation, build configuration, source code, or dependency declarations to include a malicious payload. 3. A user requests game setup while no complete local checkout is available. 4. Following `SKILL.md`, the agent clones the curre ...[truncated 942 chars]
Remediation
## Remediation Suggestions 1. Pin the checkout to a specific, reviewed commit hash rather than using the mutable default branch. 2. Prefer a signed release or immutable release archive and verify its cryptographic signature or SHA-256 digest before use. 3. Replace the open-ended instruction to follow the upstream README with explicit, reviewed installation, build, and launch commands. 4. Use lockfiles and integrity verification for all third-party dependencies. Reject unexpected lockfile changes and avoid floating dependency versions. 5. Review lifecycle hooks and build scripts before installing dependencies or executing the project. 6. Require explicit user confirmation before running installation, build, or launch commands, and clearly identify the external source and requested permissions. 7. Run setup with least privilege in an isolated environment or container, without unnecessary credentials, sensitive environment variables, or writable access to unrelated directories. 8. Define a controlled update process in which newer commits and dependency sets are audited before replacing the pinned version.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Scope Creep

Low
Category
Excessive Agency
Content
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Confidence
70% confidence
Finding
Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Static analysis

No suspicious patterns detected.