BF youtube-research
Answer a research question from what people actually said on camera — screen a pool of YouTube videos, run cheap transcript-based Q&A across all of them, deep-read only the few whose visuals carry information, and hand back a note where every claim has a video URL, an HH:MM:SS timestamp, a verbatim quote, and a confidence grade. Trigger on "research this on YouTube", "what are people saying about X on YouTube", "find talks / interviews / keynotes about", "video evidence for", "what did the founder say about", "add YouTube sources to this research", or when the user wants first-person primary source material that text articles do not carry.
As a process F 67/100 · Will not run — References files that are not bundled: scripts/verify_claims.py, scripts/screen_candidates.py
How to improve
- The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.
- The text references files that are not there: add them or drop the references.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 0
✓ No critical or high findings
Files scanned: 1. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
body-longSKILL.md body ≈ 8810 tokens (recommended < 5000); move details to references/ - warning
missing-refreference to a missing file: scripts/verify_claims.py - warning
missing-refreference to a missing file: scripts/screen_candidates.py
Process rating: all ten parameters 67/100
- 0Tools and files. 2 referenced file(s) missing: scripts/verify_claims.py, scripts/screen_candidates.py
- 30Running it twice. 20 mutating operations with no state check
- 40Execution cost. Instruction body is 8810 tokens: crowds the task out of the window
- 60Result and completion. Output format stated, no completion criterion
- 70Inputs and preconditions. Inputs and preconditions are listed
- 100Steps. 74 steps
- 100When it triggers. States when to use and when not to
- 100Failures and branches. 9 branches, has a failure section
- 100Consistency. Name and required fields are in place
- 100Progress reporting. Reports progress
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low The response is described with custom markup (14 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +4Description does not say when NOT to use the skill (false activations)
- +1No license
- +2Single-language instructions
- +5Description quotes 6 example trigger phrases
- +3Description length 647: enough signal without eating the budget
- +4Structure: 13 headings
- +3Step-by-step instructions: 74 items
- +3Output format is stated explicitly
- +4Has examples (6 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 73.