ClawHub

Meeting Transcripts

Security checks across malware telemetry and agentic risk

Overview

The skill matches its Fireflies transcript-capture purpose, but it needs review because it persists sensitive meeting records and its optional webhook can be exposed without required authentication.

Install only if you want Fireflies meeting content saved into OpenClaw memory. Prefer polling unless you need real-time webhooks, and configure a strong webhook secret before exposing the server publicly. Review stored meeting files regularly and remove the API key, cron job, or webhook tunnel when you no longer need capture.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (7)

Lp3

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding
The skill instructs use of network access and local secret storage but does not declare permissions, which weakens transparency and policy enforcement around what the skill can access. In a system that relies on declared permissions for review or sandboxing, this can lead to overbroad execution and make sensitive transcript retrieval harder to govern.

Tp4

High
Category
MCP Tool Poisoning
Confidence
84% confidence
Finding
The documented behavior does not fully match what the skill appears to do: it adds manual fetching/listing workflows, persists processed IDs locally, exposes an HTTP webhook server, and overstates extraction of a dedicated decisions section. Behavior mismatches are dangerous because operators may approve or invoke the skill without understanding its full attack surface, especially the exposed network listener and additional local persistence.

Vague Triggers

Medium
Confidence
78% confidence
Finding
The trigger guidance is broad enough to match many ordinary meeting-related requests, which can cause the skill to activate in situations where the user did not intend transcript collection or review. In this context that is risky because the skill handles sensitive meeting content and may fetch or store transcripts automatically.

Missing User Warnings

High
Confidence
95% confidence
Finding
The skill description says it writes structured markdown to memory but does not clearly warn that meeting transcripts, participant identities, summaries, and action items may contain highly sensitive personal or business information. Without an explicit warning and consent boundary, users may unknowingly persist confidential content into longer-term memory, increasing privacy, compliance, and insider access risks.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The script automatically writes fetched meeting transcripts, summaries, participants, organizer email, and full transcript text to a local memory directory without any consent prompt, warning, minimization, or access-control check. In this skill context, the data is inherently sensitive business or personal meeting content, so silent persistence materially increases confidentiality and retention risk if the workstation, workspace, backups, or downstream tools are compromised.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
When given a meeting ID, the script fetches and saves the full transcript and metadata directly to disk, again without an explicit warning or confirmation about local storage of potentially sensitive content. Because this skill is specifically designed to capture meeting transcripts, the context makes the issue more serious: the stored data may contain confidential discussions, action items, attendee identities, and other sensitive records.

Session Persistence

Medium
Category
Rogue Agent
Content
# Meeting Transcripts (Fireflies.ai)

Auto-capture meeting transcripts from Fireflies.ai, extract action items and decisions, write structured markdown to memory.

## Setup
Confidence
93% confidence
Finding
write structured markdown to memory. ## Setup ### 1. Store API key ```bash echo "YOUR_FIREFLIES_API_KEY" > ~/.openclaw/secrets/fireflies-api-key.txt ``` Get your key from Fireflies dashboard → Setti

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal