Live survey304 models · 14 providers · 22,792 auditsAudited grill-with-docs · SAFE →What changedHow we verifyllms.txt

Agent skill · Audited from source

Kb Yt Channel

skill · pedronauck/skillsLicence not declaredAudited 16 Oct 2026 · source 118e61cf1c61

★ 637pushed 31 days ago
✓SAFEGrade B · 89/100 · 16 Oct 2026 · 118e61c
637repo stars · whole pedronauck/skills

Is it safe? Yes — SAFE, grade B (89/100). We read this skill's source at commit 118e61cf1c61 and observed filesystem access and no hard-coded secrets. What it does: Create and maintain Karpathy KB topics from YouTube channels through scaffolding, bulk transcript ingestion, validation, and indexing. Excludes single videos, general summaries, and other sources.

Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
More
Cite

Cite this page

ModelsAtlas. “Kb Yt Channel (pedronauck/skills)”. Retrieved 16 Oct 2026. https://modelsatlas.com/skills/pedronauck/kb-yt-channel

Trust audit

Audit 0.4.1 · source read 16 Oct 2026

Trust score89/100✓SAFE B
0 critical0 high0 medium2 low0 info

SAFE means nothing in the source contradicts what the skill says it does. Grade B is shared by 16,667 of 17,602 audited skills.

What the source does

  • FilesystemDeclared2 observations
  • NetworkNone observedNo outbound calls in the source
  • ShellNone observedNo commands run in the source
  • DependenciesPinnedNothing floats
  • SecretsNone foundNo keys in source

Five audit layers

LayerWhat we checkResultFindings
L0 ProvenanceRepo owner, licence, archive status, activityPASS0
L1 Static sourceSecrets, path traversal, supply chain, hidden codePASS2
L2 InstructionsPrompt injection, authority framing, hidden directivesPASS0
L3 InventoryThe files and declarations match the sourcePASS0
L4 SandboxRun in isolation and watch behaviourSKIPPED why0

Gates applied: no behavioural pass.

Findings (2)

LOWFilesystem / pathfs.traversal · scripts/ingest-channel.py:212
F-05 Filesystem / pathCWE-22 · CWE-59Layer L1
text = text.replace("[root CLAUDE.md](../CLAUDE.md)", "[root CLAUDE.md](../../CLAUDE.md)")
Why
No further detail recorded for this rule.
Fix
—
LOWFilesystem / pathfs.traversal · scripts/ingest-channel.py:213
F-05 Filesystem / pathCWE-22 · CWE-59Layer L1
text = text.replace("../.claude/skills/karpathy-kb/SKILL.md", "../../.claude/skills/kb/SKILL.md")
Why
No further detail recorded for this rule.
Fix
—

Audit 0.4.1 · source sha 118e61cf1c61 · 16 Oct 2026Report an issue or request a re-scan

Provenance

Who published it, and what the verdict is pinned to

  • Publisherpedronauck GitHub userOwner of the repository this skill is read from
  • Source folderpedronauck/skills/skills/mine/kb-yt-channelThe exact folder we audited
  • LicenceNot declaredAs the repository states it
  • Commit pinned for this verdict118e61cf1c61A new audit runs when the source changes
  • Repository statusNot archivedLast push 31 days ago
  • Sandboxed runNot runSkills are read, not executed; layer L4 is skipped

What it tells the agent

Full source on GitHub ↗

Its description: Create and maintain Karpathy KB topics from YouTube channels through scaffolding, bulk transcript ingestion, validation, and indexing. Excludes single videos, general summaries, and other sources.

---
name: kb-yt-channel
description: "Create and maintain Karpathy KB topics from YouTube channels through scaffolding, bulk transcript ingestion, validation, and indexing. Excludes single videos, general summaries, and other sources."
disable-model-invocation: true
---
# KB YouTube Channel

Use the native `kb ingest channel` for channel resolution, transcripts, dedup/resume, retries, and pacing. The bundled `scripts/ingest-channel.py` adds topic organization under `yt-channels/`, metadata, indexes, validation, and a run report.

1. Resolve the vault, channel/playlist URL, topic slug/title/domain, video selection (`--limit N` or `--all`), and transcript policy (`captions`, `auto`, or `stt`) from the request. Ask only for choices that cannot reasonably be inferred.
2. Check required tools once, reusing current evidence: `kb`, `yt-dlp`, `qmd` when indexing, and `ffmpeg` plus the configured provider when STT is requested. Do not install optional networking dependencies as a default preflight.
3. Run `python3 <kb-yt-channel-dir>/scripts/ingest-channel.py` with `--vault`, `--channel-url`, `--topic-slug`, `--title`, `--domain`, selection, and `--transcribe`. Resolve the helper from this skill directory. Native-language captions are the default; override `--sub-langs` only when needed. `--no-index` defers indexing; `--embed` requests embeddings.
4. A preview is optional. The existing `--dry-run` **creates/updates the topic skeleton and metadata**, then lists the ingest plan without transcripts; it is not a read-only preview. Use it only within authorized topic creation/update. Do not require preview followed by a repeated full preflight.
5. Inspect the helper's JSON summary and failures. Reuse its completed topic-info/lint/index validation; run additional checks only for missing evidence. When searchability is part of the task, verify one representative `kb search` in the topic/collection after indexing.
6. Report actual ingested/skipped/failed counts and the report path. Compile wiki articles only if the task includes synthesis; transcript ingestion alone does not imply compilation.

## Failures and references

Start with the existing network configuration and native bounded retries/backoff. On rate limits, honor server guidance, lower concurrency, or increase pacing. Configure authenticated cookies or a permitted proxy only when an observed access/network problem requires it; never expose credentials or mandate a residential proxy for large channels.

Read `references/channel-topic-contract.md` for metadata/output details and `references/troubleshooting.md` for an actual failure. Resume the same request after repair; native dedup preserves successful ingests. Missing captions can justify `auto`/STT if that scope and provider cost are authorized. Preserve partial outputs; use the supported refresh path for a specific transcript rather than deleting user data by default.

Excerpt, first 2,916 characters of SKILL.md, verbatim from the audited commit. It is text from someone else’s repository, quoted so you can judge it before your agent reads it.

Install

Pinned to the audited commit 118e61c

Download folder ↗
git clone https://github.com/pedronauck/skills
git -C skills checkout 118e61cf1c61

The author documents no install command, so this uses the repository directly. Our verdict covers commit 118e61cf1c61 only.

Works in

The author has not said which agents this skill supports, and we have not tested it in any.

Audit history

Every audit, every source change

  1. ✓SAFE B 89 / 100

    first audit · source 118e61cf1c61

Similar skills

Questions about Kb Yt Channel

What does the Kb Yt Channel skill do?
My personal favorite skills
Is Kb Yt Channel safe to install?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.
What can Kb Yt Channel access on my machine?
The audit observed that it reads or writes files. Each of those is consistent with what it says it does. Secrets in the source: none found.
How current is this page?
The grade is for one exact copy of the source (118e61cf1c61), read on 2026-10-16. The repository is watched, and a new audit runs when it changes — this is the first audit.

Sources

Spotted something wrong? Report a correction · Methodology

Keep going in your AI

Opens with this page as the source. We add nothing else to the prompt.