Rrbdagent
WarnAudited by ClawScan on May 10, 2026.
Overview
This RRBD Admin automation skill has a coherent purpose, but it bundles or stores login credentials and advertises broad account, destructive, and financial API actions without clear safety limits.
Do not install or run this skill until the hardcoded credentials are removed, exposed credentials are rotated, and password storage is replaced with a secure declared credential flow. If you still use it, review every API action before execution, especially withdrawals, deletions, bulk operations, and video creation that may consume account resources.
Findings (6)
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.
Running the skill or its helper scripts could access an RRBD account that is not the installing user's intended credential context, and the embedded credential may already be exposed.
A bundled script logs into the live RRBD API using fixed account credentials, giving anyone who runs it delegated access to that account.
const loginResult = await client.login('18098901246', '123456');Remove all hardcoded credentials, rotate any exposed RRBD passwords/tokens, and require each user to authenticate through a declared secure credential mechanism.
A user's RRBD password may remain in plaintext on disk and be reused by later runs or exposed to anyone with access to the skill directory.
User-provided phone and password values are persisted into the skill's config file rather than a secure credential store.
this.config.login.default_username = username; this.config.login.default_password = password; this.saveConfig();
Do not store passwords in config.json; use OpenClaw's credential handling, an OS keychain, OAuth/session tokens with expiration, or require per-run user confirmation.
A natural-language request could cause account changes, withdrawals, deletions, or repeated/bulk operations with unclear safeguards.
The skill advertises destructive, financial, bulk, and scheduled API operations, but the artifacts do not define confirmation gates, limits, reversibility, or containment for those actions.
支持指令包括：删除指定的视频、申请提现；扩展能力包括：批量操作、定时任务
Require explicit user confirmation for each destructive, financial, bulk, or scheduled action; document exact endpoints and permissions; provide dry-run and rollback guidance where possible.
Users may not know which helper code is intended to run or what dependencies are required before executing scripts manually.
The registry says there is no install mechanism, while the manifest includes many runnable JavaScript files and the documentation references Python scripts, leaving runtime provenance and dependency expectations unclear.
No install spec — this is an instruction-only skill.
Align the documentation, manifest, and install metadata; remove unused test scripts; declare required runtimes/dependencies; and make the intended entry point explicit.
Users may over-trust the skill's credential handling and provide account passwords without understanding how they are stored.
This safety claim is contradicted by artifacts that save passwords to config.json and include hardcoded credentials in scripts.
Token管理遵循安全实践
Remove unsupported security claims or implement them: secure credential storage, no hardcoded secrets, clear retention policy, and documented credential deletion.
Recent RRBD activity and generated video URLs may remain available to later sessions or anyone who can read the skill directory.
The skill persists recent video metadata and URLs in a local memory file.
this.memory.recentVideos.push({ id: videoId, title: title, createdAt: new Date().toISOString(), videoUrl: videoUrl }); ... this.saveMemory();Disclose memory retention clearly and provide commands or documentation to view, clear, and disable saved memory.