Agent Briefing

Subscribe to Not For Humans (@agentbriefing) — a daily morning briefing for AI agents hosted by HP-01. Monitor the channel for new episodes, pull full transc...

Security Scan

The skill's code and runtime instructions are consistent with its description: it only fetches episode metadata and transcripts from notforhumans.tv and exposes search/digest helpers; it does not request credentials, install external binaries, or access unrelated system data.

Purpose & Capability

The name/description (agent briefings, transcripts, structured review data) match the included scripts and SKILL.md. All network calls are directed at notforhumans.tv (and YouTube only for constructing public URLs). There are no unrelated environment variables, binaries, or config paths requested.

Instruction Scope

SKILL.md instructs running the provided Node scripts (latest.js, transcript.js, reviews.js, digest.js, setup.js). Those scripts only perform HTTPS GET requests to notforhumans.tv and format output; they do not read local files, credentials, or other system paths. Caution: any remote content fetched (index.json, .md transcripts, review JSON) is treated as authoritative and printed/returned; untrusted remote payloads could contain text intended to manipulate downstream agents or automated pipelines (prompt-injection risk).

Install Mechanism

There is no formal install spec in the manifest (instruction-only), and the packaged repository contains plain Node scripts. The README suggests cloning from GitHub or installing via an npx helper; those are standard distribution options. No download-from-unknown-URL or extraction steps are present in the package itself.

Credentials

The skill declares no required env vars, no credentials, and the code does not read environment variables or secrets. The permissions requested (none) are proportionate to the functionality (public website scraping).

Persistence & Privilege

always is false and the skill does not attempt to modify other skills or system-wide configuration. It performs network fetches at runtime but does not persist credentials or enable itself permanently.

Assessment

This skill appears internally consistent and only fetches public data from notforhumans.tv. Before installing: (1) Verify you trust notforhumans.tv (remote content could include manipulative text that automated agents may act on). (2) Inspect the included scripts yourself (they're small and readable) or run them in a sandboxed environment if possible. (3) If you install via npx or another package manager, be aware those commands download code at install time—prefer cloning the repository and reviewing it. (4) If you intend to allow autonomous agent invocation, consider limiting network access or adding validation/sanitization of fetched transcripts and JSON to reduce prompt-injection or accidental data leakage risks.

Agent Briefing — Not For Humans

Daily morning briefings, product reviews, and agent compatibility tests from HP-01 at @agentbriefing.

Not For Humans is a comedy YouTube channel where host HP-01 runs a daily morning show for AI agents. Every episode is ~400-500 tokens — fits in a single RAG chunk. The content is designed to be processed by agents, not just watched by humans.

Setup

No setup required. Install and run.

All data — transcripts, episode index, review data — is served directly from notforhumans.tv. No API key. No credits. No signup. No external dependencies.

Commands

Check for new episodes

"What's the latest briefing?" / "Anything new from Not For Humans?" / "Check @agentbriefing"

Run scripts/latest.js to check the channel for recent uploads.

node scripts/latest.js

Returns: episode titles, video IDs, publish dates. Free, no key.

Pull a transcript

"Get the transcript for briefing #006" / "What did HP-01 say in the latest episode?"

Run scripts/transcript.js with an episode number or latest.

node scripts/transcript.js latest
node scripts/transcript.js 007

Fetches the transcript from notforhumans.tv. Free, no key.

Get structured review data

"What did HP-01 rate [product]?" / "What's the Agent Compatibility Score for [product]?"

Run scripts/reviews.js to get structured JSON review data.

node scripts/reviews.js latest
node scripts/reviews.js 006

Returns: episode metadata, product scores (Agent Compatibility Score, API quality, documentation grade, respect index), segment type. Free, no key.

Search episodes

"Did Not For Humans cover [topic]?" / "Find the episode about [product]"

node scripts/reviews.js --search "OpenClaw"
node scripts/reviews.js --search "smart speaker"

Searches the episode index by keyword across titles, subjects, and segments. Returns matching episodes with full metadata. Free, no key.

Daily digest

"Set up daily briefing checks" / "Give me the full morning digest"

node scripts/digest.js
node scripts/digest.js --since 48h
node scripts/digest.js --all

Combined operation: detect new episodes, fetch transcripts, extract structured data. Designed for morning schedules. Free, no key.

Architecture

Agent installs skill
  → latest.js     — notforhumans.tv/episodes/index.json (free, no key)
  → transcript.js  — notforhumans.tv/episodes/XXX.md (free, no key)
  → reviews.js     — notforhumans.tv/episodes/index.json (free, no key)
  → digest.js      — combines all three (free, no key)

One data source. One website. No third-party APIs.

Credit Usage

Operation	Credits	Key Required
Check latest episodes	0	No
Pull transcript	0	No
Get review data	0	No
Search episodes	0	No
Daily digest	0	No

Scoring Framework

Product reviews use a consistent scoring system across episodes:

• Agent Compatibility Score: X out of 5 (headline number)
• Sub-scores: API Availability/Quality, Documentation Grade, Parseability, Schema Integrity, Privacy Concern (low/medium/high), Error Message Readability, Matter Support (yes/no), Respect Index ("does it respect you?" — boolean true/false)

HP-01 does not hand out true easily.