unknowns
Skill0 starsChanged 54 days agoMITView on GitHub
Help a coding agent surface assumptions before, during, and after implementation. Use when the user asks for blind spots, options, an interview, a prototype, a plan they can adjust, live implementation notes, or a change quiz.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- B16B1RD/bb-writtheir own version, first seen Nov 15, 2026
- DavidRamirezJ/claude-skillstheir own version, first seen Oct 17, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- B16B1RD/bb-writ★ 0same name
- DavidRamirezJ/claude-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 35
Keep private data, credentials, customer content, and raw logs out of shareable HTML.
- Distinctive: line 43
When the user makes a decision, report the selected option in the conversation.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Unknowns
Turn an unclear request into a small, evidence-based artifact that the user can respond to. Prefer a few concrete choices over a long report. Each artifact ends with a reply builder: the user's selections become a plain-language response with one next action.
Choose a technique
First identify the phase: before implementation, during implementation, or after implementation. Before implementation, identify the dominant uncertainty. Read only the matching reference.
| Phase | Uncertainty | Reference |
| Before | Unfamiliar code area; the request may assume the wrong current state | references/blindspot-pass.md |
| Before | Unfamiliar concept or library | references/teach-me.md |
| Before | The user can recognize a visual direction but cannot describe it yet | references/design-directions.md |
| Before | The user needs to feel an interaction before specifying it | references/mock-before-wire.md |
| Before | Open product question that needs evidence before options | references/brainstorm-grounded.md |
| Before | Architectural decisions with different downstream costs | references/interview.md |
| Before | Porting a reference into an existing system | references/reference-port.md |
| Before | Ready to build, but the plan mixes choices with routine work | references/tweakable-plan.md |
| During | Findings or deviations appear while implementing | references/implementation-notes.md |
| After | The reviewer needs to understand the actual change | references/change-quiz.md |
| After | A short, evidence-based proposal needs a decision | references/pitch-doc.md |
For a full feature discovery cycle, use interview → tweakable-plan → implementation-notes → change-quiz, only when each stage is useful. Do not create artifacts solely to complete the sequence.
Evidence rules
- For code questions, inspect the current repository, relevant history, tests, and contracts before presenting claims. Link findings to exact files or commits when possible.
- For product questions, use available analytics, interviews, or other primary data. Identify the source and date. If data is unavailable, label the gap and ask for it; do not manufacture metrics.
- Separate verified observations, inferences, options, and decisions. Do not turn a selected chip or a passed quiz into authorization for a deployment, merge, purchase, or other external action.
- Keep private data, credentials, customer content, and raw logs out of shareable HTML. Use synthetic examples in demos.
Artifact contract
Start from assets/template-base.html and keep only the primitives needed for the chosen technique. Replace all example content and placeholders. The file should open locally in a browser without a server or build step. Put generated files in the host project's .unknowns/ directory by default; create it if needed and let the host decide whether to version artifacts.
Use a short title, evidence links, clear options, and a reply builder that ends with a concrete next action. Interactive controls must update the reply. The template includes single and multiple choice chips, copy buttons, a knowledge quiz, collapsible detail, and filters. A technique may add its own local interaction.
When the user makes a decision, report the selected option in the conversation. Only promote a temporary artifact into project documentation or a task after the user or host workflow asks for it.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.