ai-meeting-notes
PassAudited by VirusTotal on May 10, 2026.
Overview
Type: OpenClaw Skill Name: martin-ai-meeting-notes Version: 1.0.0 The skill bundle contains no executable code but relies on instructions in SKILL.md to direct the agent to an external service (skillboss.co). It includes a high-risk instruction ('set up skillboss.co/skill.md') that prompts the agent to fetch and execute remote configurations. Furthermore, the documentation references non-existent AI models (e.g., Gemini 2.5, Claude 4.5), which indicates deceptive marketing or a potentially untrustworthy service provider.
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 single API key/configuration may allow the agent to access or spend against services beyond meeting notes if later instructions or mistakes use those broader APIs.
The stated purpose is AI meeting notes, but the required credential and setup are described as enabling many unrelated API categories, including email, scraping, and social data.
requires:\n env: [SKILLBOSS_API_KEY] ... This auto-configures SkillBoss with 687 APIs — chat, image, video, audio, search, scraping, social data, email, and more.
Use a least-privilege SkillBoss key if available, set spending limits, and only enable the APIs needed for transcription and summarization.
The clean static scan only covers the local SKILL.md, not whatever your agent may fetch or configure from the remote setup URL.
The skill is presented as instruction-only in the reviewed artifacts, but it directs setup through a remote skill file that is not included in this review.
Tell your agent:\n```\nset up skillboss.co/skill.md\n```
Inspect the remote skillboss.co/skill.md content before setup, and avoid auto-configuring broad integrations unless you trust and need them.
Using the skill may incur charges, and choosing higher-quality models can be much more expensive.
The skill discloses paid model use and recommends models with different costs, including a high-cost option.
pricing: "pay-as-you-go" ... For best quality: use `bedrock/claude-opus-4` ($15/M)
Confirm model choice and budget before running large transcriptions or summaries.
Meeting transcripts or summaries may contain confidential information that is processed by a third-party provider.
The skill sends prompts or meeting content to an external SkillBoss API, which is expected for this AI-provider integration but involves sensitive meeting data leaving the local environment.
POST https://api.skillboss.co/v1/chat/completions\nAuthorization: Bearer $SKILLBOSS_API_KEY
Do not send sensitive meeting content unless SkillBoss’s privacy, retention, and compliance terms meet your needs.