From unbuilt
Use BEFORE adding a new feature, expanding scope, adding a dependency, re-introducing something that was previously removed, or reversing a "keep it simple" decision. Reads the project's UNBUILT.md and warns if the planned change conflicts with something that was deliberately left unbuilt, so deliberate decisions are not reversed by accident. Also runs on request to audit a plan against the ledger.
How this skill is triggered — by the user, by Claude, or both
Slash command
/unbuilt:checkThese tools are removed from Claude's available pool while this skill is active:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Before scope grows, make sure you are not about to undo a deliberate decision.
Before scope grows, make sure you are not about to undo a deliberate decision.
Find the ledger. Look for UNBUILT.md in the project root.
Read it and compare the change you are about to make against each Left unbuilt entry. Match on the intent or capability described, not just shared words — a plan to "add Redis for response storage" conflicts with an entry that ruled out "a caching layer." When you are genuinely unsure whether something matches, surface it briefly and let the user decide rather than staying silent.
If the planned change matches or closely relates to a recorded entry, pause before proceeding and tell the user plainly:
Then ask whether to:
Do not silently cross a recorded boundary.
If there is no conflict, continue with the task. Only mention the check if it is relevant — don't add noise.
UNBUILT.md (recording is the other
skill's job).npx claudepluginhub gru-953/unbuilt --plugin unbuiltProvides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.