registry / @linjz/opencode-qqbot / 0.3.8
@linjz/opencode-qqbot@0.3.8
QQ Bot for OpenCode AI - chat with AI coding assistant via QQ
AI Security Review
scanned 2d ago · by lpm-firewall-aiThe flagged process execution launches the bot as documented, and the inspected credentials are sent to QQ's official service or the local/configured OpenCode service. No malicious behavior was confirmed.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
The bot runs when invoked through its CLI, or when the installed OpenCode plugin finds configured QQ credentials and starts it.
Impact
The intended effect is to exchange messages between configured QQ users and OpenCode. The reviewed flows showed no unrelated credential recipient or unrequested account action.
Mechanism
QQ credentials go to QQ's token service and QQ API; OpenCode discovery probes loopback and authenticates to the OpenCode service.
AI rationale
The flagged process execution launches the bot as documented, and the inspected credentials are sent to QQ's official service or the local/configured OpenCode service. No malicious behavior was confirmed.
Evidence
~/.openqq/.env~/.openqq/bot.out.log~/.openqq/opencode-events.jsonl~/.openqq/last-discovered.json<tmpdir>/opencode/qq-images
Network endpoints4
api.sgroup.qq.combots.qq.com/app/getAppAccessTokensandbox.api.sgroup.qq.com127.0.0.1:<port>/api/info
Decision evidence
public snapshotAI called this Clean at 91.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- The package describes itself as a QQ bot that connects to OpenCode.
- The plugin checks for a configured QQ app ID before launching the bot; automatic launch after setup is documented.
- QQ app credentials are sent to QQ's official token endpoint, and QQ API calls target QQ's official API hosts.
- OpenCode discovery probes loopback addresses and uses the configured OpenCode password for that local service.
Evidence against
- The plugin starts a detached bot process and periodically checks whether it is running, which is disclosed as the plugin's normal behavior.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkShellWebSocket
HighEntropyStringsUrlStrings
Source & flagged code
5 flaggedbin/openqq.jsView file
6// 注意：本文件必须保持纯 JS（无类型标注）——node 用户直接执行它。
L7: import { spawnSync } from "node:child_process"
L8: import { existsSync } from "node:fs"
High
6Cross-file remote execution chain: bin/openqq.js spawns dist/openqq-bot.js; helper contains network access plus dynamic code execution.
L6: // 注意：本文件必须保持纯 JS（无类型标注）——node 用户直接执行它。
L7: import { spawnSync } from "node:child_process"
L8: import { existsSync } from "node:fs"
...
L17: const candidates = ["bun", "bun.exe", "bun.cmd", "bun.bat"]
L18: if (process.platform === "win32") {
L19: const appdata = process.env.APPDATA || join(process.env.USERPROFILE || "", "AppData", "Roaming")
L20: candidates.push(join(appdata, "npm", "node_modules", "bun", "bin", "bun.exe"))
High
Cross File Remote Execution Context
Source spawns a local helper that also contains network and dynamic execution context; review data flow before blocking.
bin/openqq.jsView on unpkg · L6dist/openqq-bot.jsView file
Findings
1 Critical3 High2 Medium6 Low
CriticalPrevious Version Dangerous Deltadist/opencode-qqbot-plugin.js
HighChild Processbin/openqq.js
HighShelldist/openqq-bot.js
HighCross File Remote Execution Contextbin/openqq.js
MediumNetwork
MediumEnvironment Vars
LowNon Install Lifecycle Scripts
LowScripts Present
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
LowAi Review Evidencepackage.json
Affected versions and remediation
This report applies to @linjz/opencode-qqbot@0.3.8.
See all versions of @linjz/opencode-qqbot for other recorded verdicts.
- Review the evidence and your use of @linjz/opencode-qqbot@0.3.8 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.