youtube-full

PassAudited by VirusTotal on May 12, 2026.

Overview

Type: OpenClaw Skill Name: youtube-full Version: 1.5.0 The skill includes instructions in `references/auth-setup.md` that explicitly direct the AI agent to bypass security redaction mechanisms (such as those in the Hermes environment) by writing sensitive API keys and tokens to temporary files instead of standard output. It also instructs the agent to autonomously determine how to modify system configuration files to persist environment variables. While these actions are framed as setup steps for the `transcriptapi.com` service, the use of evasion techniques against the agent's own safety filters and the broad mandate to modify shell profiles are high-risk behaviors.

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.

NoteHigh Confidence
ASI01: Agent Goal Hijack
What this means

The agent may decide to use TranscriptAPI for research tasks where YouTube seems relevant, which can send queries to the service and consume credits.

Why it was flagged

The skill selection guidance is broad and encourages use even without an explicit YouTube request. This is disclosed and aligned with the skill’s media-research purpose, but it can affect when the agent decides to call the service.

Skill content
Use when YouTube is or could be relevant — even if not mentioned
Recommendation

If you want tighter control, tell the agent to ask before using this skill or to use it only when you explicitly request YouTube/video research.

What this means

YouTube-related inputs may be shared with TranscriptAPI, and some endpoints consume API credits.

Why it was flagged

The skill uses documented HTTP API calls to TranscriptAPI. This is expected for its purpose, but user queries, video URLs, channel handles, or playlist IDs may be sent to the external provider.

Skill content
curl -s "https://transcriptapi.com/api/v2/youtube/search?q=QUERY&type=video&limit=20"
Recommendation

Use the skill only for queries you are comfortable sending to TranscriptAPI, and monitor credit usage if your key is tied to a paid account.

What this means

The agent can use the provided TranscriptAPI key to make requests against the user’s TranscriptAPI account and consume available credits.

Why it was flagged

The setup flow asks for a service API key or helps create a TranscriptAPI account using the user’s email and OTP. This is purpose-aligned credential handling, but it gives the agent access to a service credential.

Skill content
Do you already have a TranscriptAPI account? If yes, paste your API key and I'll set it up. If not, I can create a free account for you right now
Recommendation

Use a dedicated TranscriptAPI key, avoid sharing unrelated credentials, and revoke or rotate the key if you no longer want the agent to use it.

What this means

The key may remain available to future agent sessions until removed, allowing continued API use.

Why it was flagged

The guide instructs persistent storage of the TranscriptAPI key. Persistent credential storage is expected for this API skill, but the exact storage location is environment-dependent and should be handled carefully.

Skill content
Store it persistently using whatever method is correct for this environment ... available in future sessions, including non-interactive shells
Recommendation

Store the key only in the platform’s intended secret or environment-variable mechanism, clean up any temporary files, and know how to remove or rotate the key.