Telegram Chat Summary

Security checks across static analysis, malware telemetry, and agentic risk

Overview

This looks like a real Telegram summarizer, but setup can change global OpenClaw tool permissions and the plugin handles Telegram sessions/messages on a background schedule.

Install only if you trust this plugin with your Telegram account and chat contents. Before use, inspect the flagged `src/telegram-client.ts` password line, review the OpenClaw config changes made by setup, restrict the tool allowlist/profile where possible, protect `apiHash`, `sessionString`, and bot tokens, and confirm the scheduler and summary destination are exactly what you intend.

SkillSpector

By NVIDIA

SkillSpector has not run for this release. Legacy ClawScan findings remain available under Risk analysis.

Static analysis

Exposed secret literal

Critical

Finding: File appears to expose a hardcoded API secret or token.
Skill content: password: [REDACTED],

VirusTotal

VirusTotal engine telemetry is currently stale for this artifact.

View on VirusTotal

Risk analysis

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.

ASI03: Identity and Privilege Abuse

Medium

What this means

Running setup may broaden or alter what tools the agent can use across OpenClaw, potentially affecting behavior beyond Telegram summaries.

Why it was flagged

Setup writes global OpenClaw tool permission settings, not just this plugin's own Telegram configuration.

Skill content

const REQUIRED_TOOLS_PROFILE = "full"; ... { path: "tools.profile", value: REQUIRED_TOOLS_PROFILE }, { path: "tools.allow", value: allowedTools },

Recommendation

Review `~/.openclaw/openclaw.json` before and after setup; keep only the Telegram summary tools you actually need, and avoid a broader global tool profile if your OpenClaw setup supports a more restricted configuration.

ASI03: Identity and Privilege Abuse

Medium

What this means

Anyone who obtains the session string or API hash may be able to abuse your Telegram account access.

Why it was flagged

The plugin requires Telegram API credentials and a Telegram session string, which can grant account-level access for listing chats, reading messages, and sending summaries.

Skill content

"apiHash": { "type": "string" ... }, "sessionString": { "type": "string" ... }, ... "apiHash": { ... "sensitive": true }, "sessionString": { ... "sensitive": true }

Recommendation

Treat `apiHash`, `sessionString`, and any bot token as secrets; store them only in trusted local config, do not commit them, and revoke/regenerate them if exposed.

ASI10: Rogue Agents

Medium

What this means

Configured Telegram chats may be read and summarized repeatedly in the background until the scheduler is disabled and the gateway is restarted.

Why it was flagged

The plugin registers a recurring cron task that continues to summarize configured chats while the gateway process is running.

Skill content

const task = cron.schedule(schedule, async () => { ... for (const chatEntry of cfg.chats ?? []) { ... await summarizeChat({ api, cfg }, chatEntry, signal); } });

Recommendation

Only configure chats you intend to monitor, confirm the schedule, and use the stop command plus gateway restart when you want monitoring to end.

ASI07: Insecure Inter-Agent Communication

Medium

What this means

Private chat content may be exposed to the configured AI runtime, summary destination, and local logs in summarized form.

Why it was flagged

Telegram message content is processed by an embedded AI agent, then summaries can be sent to a Telegram destination and logged.

Skill content

调用 `api.runtime.agent.runEmbeddedPiAgent(...)` 生成摘要。发送到 `targetChatId`（缺省为当前 chat）。记录日志（抓取条数、摘要内容、发送结果）。

Recommendation

Use this only for chats whose participants and data policy allow AI summarization; verify `targetChatId`, bot-token settings, and log retention.

ASI01: Agent Goal Hijack

Low

What this means

A message in a monitored chat could manipulate or distort the summary output.

Why it was flagged

Untrusted Telegram chat messages are inserted into the summarization prompt; malicious chat text could try to influence the generated summary.

Skill content

聊天记录：\n{messages}\n\n请提供结构清晰的摘要，使用要点形式。

Recommendation

Treat generated summaries as untrusted analysis of chat content, and consider adding prompt text that explicitly treats chat messages as data rather than instructions.