Banshee's Last Cry Game CN
v1.0.0恐怖惊魂夜 - 互动式恐怖悬疑剧本杀游戏。玩家将在雪山旅馆中经历一场惊心动魄的生存冒险，通过调查线索、做出选择来推动剧情发展。支持多结局、角色扮演、线索系统。适用于互动小说、剧本杀、文字冒险游戏等场景。
⭐ 1· 1.7k·2 current·2 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The skill's name/description match the included runtime instructions and the 20+ bundled chapter/reference files. It does not request unrelated environment variables, binaries, or external services; all required assets are present in the repository (references/). The operations (presenting narrative, updating a local game state) are proportional to a text-adventure skill.
Instruction Scope
Runtime instructions tell the agent to read the bundled references/* files and to create/update a local game_state.json to track progress, clues, trust levels, etc. This is expected for a save/stateful interactive game, but it does mean the skill will write a local JSON file in the agent working directory. The SKILL refers to an external design doc (CLAUDE.md) which is not present in the manifest — a minor inconsistency but not a security issue. There are no instructions to read arbitrary system files or call external endpoints.
Install Mechanism
No install spec or code files beyond narrative markdown files. Instruction-only skills have minimal disk footprint (only the provided manifest files). No downloads, package installs, or extracts are specified.
Credentials
The skill declares no required environment variables, credentials, or config paths. The runtime instructions reference only files bundled with the skill and a local game_state.json. No secrets or unrelated environment access are requested.
Persistence & Privilege
always:false (default) and normal autonomous invocation are used. The skill will persist game state to game_state.json in the agent's working directory — this is reasonable for a game but is the only form of persistence. The skill does not ask to modify other skills or system settings.
Assessment
This skill appears internally consistent for running an interactive horror/visual-novel style game. Before installing or enabling it, consider: (1) The skill will create and update a local file game_state.json in the agent's working directory — that file may contain player-chosen names, clues, and any typed content, so if you run the skill in an environment with sensitive data or shared logs, treat that file accordingly. (2) The SKILL references CLAUDE.md (an external design doc) which is not included; that is likely harmless but means some referenced design guidance isn't bundled. (3) Because it can run autonomously, run it in a sandboxed agent environment if you do not fully trust the source (the skill contains only narrative files and state updates; there are no network calls or credential requests, but sandboxing is good practice). (4) If you want to preserve or inspect saved games, back up game_state.json; if you don't want persistent state, run the skill in a disposable workspace. Overall there are no red flags for credential exfiltration, external downloads, or unexpected system access.Like a lobster shell, security has layers — review code before you run it.
latestvk979kxy90470st45pbb16jw92n8017an
License
MIT-0
Free to use, modify, and redistribute. No attribution required.