Blindspots

Skill btucker/blindspots/skills/blindspots

This skill should be used when the user is running blindspots testing, when a ".blindspots/" directory exists in the project, when the user mentions "blindspots", "dogfood", "user-trial", "experiment", or "interview" in a testing context, or when working in a git branch starting with "blindspots/".From its SKILL.md

Install

npx -y skills add btucker/blindspots --skill blindspots

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.2 KB, as published. Nobody here has run it

Blindspots — Find Your Product's Blind Spots

This skill provides context and guidance during an active blindspots session.

Modes

Command	Purpose	Has Specs?	Fixes Bugs?
`blindspots:setup`	Interactive config and persona generation	N/A	No
`blindspots:dogfood`	Test against specs, find regressions	Yes	Yes
`blindspots:user-trial`	Persona discovers the product with no insider knowledge	No	No
`blindspots:experiment`	A/B compare two variants with user trial personas	No	No
`blindspots:interview`	Ask personas questions about the product	Depends	No

The .blindspots/ Directory

Config is checked in; session output is gitignored.

Config (checked in)

File	Purpose
`.blindspots/config.md`	Project testing config (setup, start, explore, diagnose, test, specs)
`.blindspots/personas.md`	Pool of user personas

Output (gitignored)

File	Mode	Purpose
`.blindspots/dogfood-journals/<persona>.md`	dogfood	Per-persona session log
`.blindspots/user-trials/<persona>/`	user-trial	Per-persona trial output (journal, discovered-specs, reactions, comparison, report.html)
`.blindspots/screenshots/`	all	Evidence
`.blindspots/experiments/<name>/`	experiment	A/B test output
`.blindspots/interviews/<name>.md`	interview	Interview transcripts

Agents

user-trial-explorer-browser — Browser-based user trial explorer. No Read/Grep/Glob.
user-trial-explorer-terminal — Terminal-based user trial explorer. Bash + Read (restricted) + Write.
user-trial-compare — Analyzes discovered specs vs actual specs.
experiment-verdict — Synthesizes A/B test results into ship/no-ship verdict.
interviewer — Conducts persona Q&A sessions.

Shared Steps

All commands auto-run /setup if .blindspots/config.md is missing. The shared startup pattern:

Check for .blindspots/config.md — run blindspots:setup if missing
Check for .blindspots/personas.md — run blindspots:setup if missing
Select persona — method varies by command:
- /dogfood, /user-trial: random, or --persona <name>
- /interview: shows persona list with panel option, asks user to pick
- /experiment: prompts interactively (test name, variants, user type, run count)
Run setup commands from config
Infer interface mode from ## Start (browser or terminal)
Mode-specific launch with BLINDSPOTS_DEPTH=1

Reference Files

references/dogfood-cycle.md — Guided testing cycle (EXPLORE → USE → DIAGNOSE → FIX → COMMIT)
references/compare-prompt.md — Spec comparison methodology
references/persona-template.md — Persona generation template
references/experiment-verdict-prompt.md — A/B verdict methodology