Blindspot Pass
Neeeophytee/finding-unknowns-skills
Surface the user's unknown unknowns before work starts. An agent skill from Neeeophytee/finding-unknowns-skills.
A skill your agent uses when the user asks for a blindspot pass or to find their unknown unknowns, or signals unfamiliarity with a domain, tool, or codebase area ("never used X", "first time doing…
MITAuto-check passed
Install Blindspot
$ npx skills add vinta/hal-9000 --skill blindspot -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install vinta/hal-9000 blindspot --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/vinta/hal-9000.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/blindspot .claude/skills/blindspot && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "blindspot" agent skill from https://github.com/vinta/hal-9000/tree/main/skills/blindspot into .claude/skills/blindspot/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "blindspot", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/vinta/hal-9000/tree/main/skills/blindspotType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add vinta/hal-9000 --skill blindspot -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install vinta/hal-9000 blindspot --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/vinta/hal-9000.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/blindspot .agents/skills/blindspot && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "blindspot" agent skill from https://github.com/vinta/hal-9000/tree/main/skills/blindspot into .agents/skills/blindspot/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "blindspot", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add vinta/hal-9000 --skill blindspot -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install vinta/hal-9000 blindspot --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/vinta/hal-9000.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/blindspot .cursor/skills/blindspot && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "blindspot" agent skill from https://github.com/vinta/hal-9000/tree/main/skills/blindspot into .cursor/skills/blindspot/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "blindspot", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/vinta/hal-9000.git --path skills/blindspot--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add vinta/hal-9000 --skill blindspot -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install vinta/hal-9000 blindspot --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/vinta/hal-9000.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/blindspot .gemini/skills/blindspot && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "blindspot" agent skill from https://github.com/vinta/hal-9000/tree/main/skills/blindspot into .gemini/skills/blindspot/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "blindspot", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install vinta/hal-9000 blindspotInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add vinta/hal-9000 --skill blindspot -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/vinta/hal-9000.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/blindspot .github/skills/blindspot && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "blindspot" agent skill from https://github.com/vinta/hal-9000/tree/main/skills/blindspot into .github/skills/blindspot/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "blindspot", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add vinta/hal-9000 --skill blindspot -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install vinta/hal-9000 blindspot --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/vinta/hal-9000.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/blindspot .opencode/skills/blindspot && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "blindspot" agent skill from https://github.com/vinta/hal-9000/tree/main/skills/blindspot into .opencode/skills/blindspot/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "blindspot", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
blindspot- GitHub stars
- 138
- Token cost
- ~1.6k tokens
- SKILL.md length
- 962 words
- Files
- 1
- Skills in repo
- 15
- Repo updated
- First seen
- Licence
- MIT
At a glance
A skill your agent uses when the user asks for a blindspot pass or to find their unknown unknowns, or signals unfamiliarity with a domain, tool, or codebase area ("never used X", "first time doing…
- Works in 5 steps: Recon → Show the territory → Ask which direction → …
- The user asks for a blindspot pass
- SKILL.md covers Workflow and Constraints
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Blindspot is an agent skill from vinta/hal-9000. Use when the user asks for a blindspot pass or to find their unknown unknowns, or signals unfamiliarity with a domain, tool, or codebase area ("never used X", "first time doing Y", "no idea where to start", "don't know what I don't know") before working there. Maps the areas their request leaves unnamed, purpose and shape before mechanics, shows how people usually answer each and why, and asks which to explore next, so unknown unknowns become known unknowns they can prompt with. Recommendations on a tool already…
Its SKILL.md is about 1.6k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
The repository describes itself as: Opinionated AI coding agent and dev environment automation for macOS. The licence is MIT.
When your agent uses it
- The user asks for a blindspot pass
- Find their unknown unknowns
- Signals unfamiliarity with a domain
- Codebase area (never used X
Example prompts
- “never used X”
- “first time doing Y”
- “no idea where to start”
- “/blindspot”
Requirements
- Pre-approved tools (allowed-tools): WebSearch
Workflow steps
5 steps, taken from the step headings in SKILL.md.
What it can do on your machine
Read from SKILL.md and the folder at commit 23243bf. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves these tools, so the agent can use them without asking each time:
WebSearch
From allowed-tools in the SKILL.md frontmatter.
Runs code
No scripts in the folder and no shell commands in SKILL.md.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Blindspot loads about 1.6k tokens when it runs. Until then it costs about 140 tokens; SKILL.md has 962 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from vinta/hal-9000 at commit 23243bf, republished under its MIT licence (© vinta). 962 words, ~1,635 tokens.
Download SKILL.mdSave it as
.claude/skills/blindspot/SKILL.md (or your agent's skills folder).- name
- blindspot
- description
- Use when the user asks for a blindspot pass or to find their unknown unknowns, or signals unfamiliarity with a domain, tool, or codebase area ("never used X", "first time doing Y", "no idea where to start", "don't know what I don't know") before working there. Maps the areas their request leaves unnamed, purpose and shape before mechanics, shows how people usually answer each and why, and asks which to explore next, so unknown unknowns become known unknowns they can prompt with. Recommendations on a tool already chosen belong to best-practices
- allowed-tools
- WebSearch
- argument-hint
- [unfamiliar topic, tool, or codebase area]
Blindspot
The user is about to work in territory they don't know. Two kinds of blindness live there: unknown unknowns (questions they don't know exist) and unknown knowns (assumptions too obvious to write down, and things they're sure of that are wrong). Convert the first into known unknowns, name the second, then hand back a map they can prompt with.
Boundaries: this skill goes wide over the areas the user's request leaves unnamed, one paragraph each; best-practices goes deep on a topic the user has already named. grilling stress-tests decisions the user can defend; this skill maps territory where they can't decide yet, so it asks which direction to explore, never which option to pick. Neither a tutorial nor a plan.
Workflow
1. Recon
Facts are your job, never the user's. Sweep before writing anything:
- Goal: the goal the request serves, climbed one level at a time through the request and the repo's records until they run out. The map sits at the highest level still open; the request's own wording is usually the lowest.
- Premises: what the request and the repo's design records (ADRs, docs, past decisions) take as settled, restated as questions. Premises are the source of idea-level areas; docs and pitfall searches yield mechanics.
- Repo: the
Exploreagent in Claude Code or anexplorersubagent in Codex, for existing patterns, conventions, and adjacent solutions. - Tools and domain:
find-docsfor current APIs and config;WebSearchin Claude Code orweb_searchin Codex for pitfalls ("X gotchas", "X common mistakes") — pitfalls live in issue threads and post-mortems, not getting-started docs.
Recon hunts for the areas a practitioner would have on their list that the request never mentions. Done when every area that could reorder the map has its source, recon or a premise, or a note that recon found nothing.
2. Show the territory
Before any question, one reply the user reads, under a page:
- Framing: the assumptions the request takes for granted, and the missing information that would change the approach, as bullets. These are the unknown knowns; naming them is the point. A premise the user can settle with a word or a fact stays here; one that takes a dig to answer is an area.
- Areas: numbered, ranked by altitude, then by how much the answer would reorder the rest of the map. Altitudes, top down: what the thing is for and what "better" means; which shape it could take and who owns what; mechanics (APIs, limits, schemas, code seams, tool picks). Mechanics become areas only when nothing above them is open; until then a mechanic recon turned up is one line inside the area it would answer. Idea-level: "is the model a classifier that settles what rules can't, or an author that rewrites the text?"; mechanic: "does the provider accept a string-enum root schema?". Each area is one paragraph: the question an expert would ask here, how people usually answer it and why, and what in this repo or situation makes it bite, with its source: recon or the premise it restates. Where the request already matches usual practice, one line saying so. "No significant unknowns here" is a valid result; a manufactured concern erodes trust faster than a short map.
Show full SKILL.md (426 more words)Show less
3. Ask which direction
Use AskUserQuestion in Claude Code or request_user_input in Codex when available.
One single-select question: the top 3 unexplored areas by number, plus "Enough, hand off". Each option is an area and its description that area's why in one line; every option names an area the reply above explained. The user is choosing where to look, so labels stay unranked: no "(Recommended)". A free-text answer reaches any numbered area.
Round 1 adds one premortem question phrased in past tense — "it's three months later and this failed: what broke?" — past tense recruits prospective hindsight; "what could go wrong" is measurably weaker. Its answer often names the real goal, so re-rank the areas by it. If the user's goal or familiarity is still unclear after recon, round 1 also carries one calibration question.
"You pick" means take the top-ranked area and record the pick as a named assumption.
4. Dig, then repeat
Explore the picked area one altitude down, never two: a goal-level area opens into the shapes it could take; a shape-level area opens into what each costs to build. Only at mechanics does a tool or setup choice go to the best-practices skill; anything else gets the same recon one level down. Show what turned up in the shape of step 2 (sub-areas join the numbered map, unpicked areas stay on it, re-ranked), then ask again. Stop when the user picks "Enough", when a round surfaces nothing new, or when what's left is cheaper to learn while building — say which.
5. Hand off
End with:
- Territory map: explored areas with what each dig settled; unexplored areas, each with the usual practice as its default; named assumptions — every open point this skill resolved by guessing gets its own bullet; recon sources cited so the user can dig deeper.
- Sharpened prompt draft the user could send: explored areas resolved inline, unexplored ones listed as open questions with their defaults.
- Offers, not auto-runs: stress-test the now-visible decisions with the
grillingskill, get recommendations viabest-practices, or enter plan mode.
Every explored area lands in the map or the prompt draft — a dig that shapes nothing was a wasted round.
Constraints
- Ask only what recon can't answer: a question the codebase or docs already answer wastes a round and erodes trust.
- Teach to prompt, not to master. If a findings reply or the hand-off exceeds roughly one page, cut.
- Ground every fact in recon, not training data: a stale fact plants a false known-known. A question needs no source beyond the premise it restates.
© vinta, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
Just SKILL.md in skills/blindspot of vinta/hal-9000.
Open the folder on GitHubat commit 23243bf
Compare with similar skills
Blindspot next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Blindspot this skillvinta/hal-9000 | 138 | — | ~1.6k | Automated safety check: Pass | MIT | |
| Blindspot PassNeeeophytee/finding-unknowns-skills | 343 | — | ~482 | Automated safety check: Pass | MIT | |
| Blindspot Passjellydn/my-ai-tools | 123 | — | ~1.7k | Automated safety check: Notes | MIT | |
| Babysit PR To Pass CIsgl-project/sglang | 37k | 2 repos | ~3k | Automated safety check: Pass | Apache-2.0 | |
| Detecting Pass The Hash Attacksmukul975/Anthropic-Cybersecurity-Skills | 34k | — | ~904 | Automated safety check: Pass | Apache-2.0 | |
| Detecting Pass The Ticket Attacksmukul975/Anthropic-Cybersecurity-Skills | 34k | — | ~717 | Automated safety check: Pass | Apache-2.0 |
Similar skills
Blindspot Pass
jellydn/my-ai-tools
Inspect history and architecture for hidden constraints before risky cross-cutting changes.
Babysit PR To Pass CI
sgl-project/sglang
Start and persistently pursue a goal to babysit an SGLang pull request until selected GitHub Actions workflows pass on the latest PR head.
Detecting Pass The Hash Attacks
mukul975/Anthropic-Cybersecurity-Skills
Detect Pass-the-Hash (T1550.002) attacks by analyzing NTLM authentication patterns, flagging Type 3 logons using NTLM where Kerberos would be expected, and correlating with credential-dumping…
Detecting Pass The Ticket Attacks
mukul975/Anthropic-Cybersecurity-Skills
Detect Kerberos Pass-the-Ticket (PtT) attacks by analyzing Windows Event IDs 4768, 4769, and 4771 for anomalous ticket usage patterns, with detection queries for Splunk and Elastic SIEM.
Blindspot
scunning1975/MixtapeTools
Peripheral vision audit for empirical output. An agent skill from scunning1975/MixtapeTools.
More from vinta/hal-9000
All 15 skills in this repoPublish Plugins Version Bump
vinta/hal-9000
Finds which plugins in the repository changed, bumps only the ones not already bumped since origin/main, and checks that each plugin's two manifests stay in sync.
Claude Code Settings Audit
vinta/hal-9000
Audits your Claude Code settings and environment variables against the current official docs and returns a ranked list of changes tied to how you actually work.
Logical Git Commits
vinta/hal-9000
Commits everything in the working tree as one logical change per commit, splitting files by hunk, with bodies that say what was wrong before and never an invented reason.
PR
vinta/hal-9000
A skill your agent uses when the user explicitly asks to push the current branch and open a PR, rewrite an open PR's body from its commits, or wait for CI and merge it
Ansible Playbook Updater
vinta/hal-9000
Keeps Ansible install tasks in line with upstream docs by fixing version, install-method and link drift one tool at a time, with a commit for each.
Simple English Rewriter
vinta/hal-9000
Rewrites docs, READMEs, issues, comments or UI text in plain Global English that translates well and still sounds native, keeping every fact intact.
Questions about Blindspot
What does Blindspot do?
A skill your agent uses when the user asks for a blindspot pass or to find their unknown unknowns, or signals unfamiliarity with a domain, tool, or codebase area ("never used X", "first time doing…. Blindspot is an agent skill from vinta/hal-9000. Use when the user asks for a blindspot pass or to find their unknown unknowns, or signals unfamiliarity with a domain, tool, or codebase area ("never used X", "first time doing Y", "no idea where to start", "don't know what I don't know") before working there.
When should I use Blindspot?
Blindspot fits situations like: the user asks for a blindspot pass; find their unknown unknowns; signals unfamiliarity with a domain; codebase area (never used X.
How do I install Blindspot in Claude Code?
Run `npx skills add vinta/hal-9000 --skill blindspot -a claude-code`. Or copy the skill folder (skills/blindspot in vinta/hal-9000) into .claude/skills/blindspot in your project. Claude Code loads it when a task matches its description.
How do I install Blindspot in Codex?
Run `npx skills add vinta/hal-9000 --skill blindspot -a codex`. Or copy the skill folder (skills/blindspot in vinta/hal-9000) into .agents/skills/blindspot in your project. Codex loads it when a task matches its description.
Can I use Blindspot in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add vinta/hal-9000 --skill blindspot -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/blindspot, .gemini/skills/blindspot, .github/skills/blindspot and .opencode/skills/blindspot in your project.
What does Blindspot need to run?
SKILL.md names no scripts, command-line tools or credentials: Blindspot is instructions for the agent only. Its frontmatter pre-approves these tools: WebSearch.
Does Blindspot access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Blindspot safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Blindspot use?
Blindspot is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Blindspot use?
About 1.6k tokens (SKILL.md is roughly 6.5k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Blindspot?
Skills that share tags, products or a category with Blindspot: Blindspot Pass (Neeeophytee/finding-unknowns-skills, 343 stars), Blindspot Pass (jellydn/my-ai-tools, 123 stars), Babysit PR To Pass CI (sgl-project/sglang, 37k stars) and Detecting Pass The Hash Attacks (mukul975/Anthropic-Cybersecurity-Skills, 34k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Blindspot?
vinta (a GitHub user) maintains it in vinta/hal-9000, which has 138 GitHub stars. The repository holds 15 skills in this directory. The repository was last updated on October 7, 2026.
Source: vinta/hal-9000 on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.