dnd5e-dm-general
PendingVirusTotal audit pending.
Overview
No VirusTotal analysis has been recorded yet. File reputation checks will appear here once the artifact hash has been scanned.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
A casual in-game request could permanently change how the DM agent behaves later, including changes that accidentally weaken rules or alter future sessions.
This directs the agent to persist player-provided changes into rule/instruction files and follow them in future sessions, without documented confirmation, scope limits, review, or rollback controls.
记录玩家提出的优化意见 ... 写入对应的规则文件（检定规则.md / DM_RULES.md） ... 以后按新规则执行
Require explicit user confirmation before modifying persistent rule files, store customizations in a separate user config/change log, and prevent player text from overriding safety or core skill instructions.
The agent may inspect and change local campaign saves or character files during normal play.
The skill instructs the agent to scan, read, and modify local D&D files. This is purpose-aligned for campaign management, but users should know it is not purely conversational.
扫描 `E:\AI\DND\saves\` ... 载入存档：读取对应 `存档*.json` ... 同时更新角色卡（index.html）中的xp数值。
Use a dedicated D&D folder, keep backups of saves and character sheets, and ask the agent to confirm before overwriting important files.
Generated maps or campaign information could be sent to the wrong Feishu chat if the account or recipient is not clearly controlled.
The skill expects generated map files to be sent through Feishu. If the agent has Feishu access, this uses a delegated communication account, but the recipient/channel approval process is not specified.
格式标准：.xlsx 文件，通过飞书发送 ... 保存后通过飞书重新发送
Confirm the Feishu recipient/channel before enabling sends, require per-send approval for shared files, or disable Feishu tools when running the skill locally.