set-up-stack-profile
Skill7 starsChanged 15 days agoMITView on GitHub
Use when starting in a repo with no .claude/stack-profile.md, when the stack changed, or when the user asks which plugins and skills to use. Detects the stack, asks only the gaps, writes the profile, and lists plugins and skill order.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- VelimirMueller/claude-skillsword for word, first seen Oct 11, 2026
Other versions1 project uses a version of this.
Word for word · 1
- VelimirMueller/claude-skills★ 7100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
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.
In this repository156 other files it gives its agents.
- skills/swift/swiftui-design-principles/AGENTS.mdAGENTS.md
- skills/swift/ios-code-audit/CLAUDE.mdCLAUDE.md
- skills/swift/swift-focusengine-pro/llms.txtllms.txt
- build-llm-seamSkill
- build-mcp-serverSkill
- build-rag-pipelineSkill
- secure-llm-featuresSkill
- set-up-llm-evalsSkill
- build-nextjs-backendSkill
- build-supabase-edge-functionSkill
- configure-backend-testsSkill
- design-http-apiSkill
- harden-backendSkill
- scaffold-fastapi-serviceSkill
- scaffold-go-serviceSkill
- scaffold-hono-serviceSkill
- secure-supabase-rlsSkill
- set-up-backend-authSkill
- set-up-background-jobsSkill
- set-up-databaseSkill
- set-up-nextjs-supabase-authSkill
- set-up-observabilitySkill
- set-up-supabaseSkill
- build-cliSkill
- release-cliSkill
- set-up-dev-toolchainSkill
- audit-securitySkill
- audit-toolchainSkill
- extend-skillsetSkill
- promote-releaseSkill
- write-commit-messagesSkill
- write-pull-requestsSkill
- analyze-complexitySkill
- apply-algorithmsSkill
- apply-design-patternsSkill
- apply-design-principlesSkill
- apply-solidSkill
- choose-data-structureSkill
- debug-systematicallySkill
- enforce-dependency-ruleSkill
- refactor-safelySkill
- slice-and-estimateSkill
- write-clean-testsSkill
- write-readable-codeSkill
- clean-frontend-scaffoldingSkill
- configure-accessibilitySkill
- configure-analyticsSkill
- configure-ciSkill
- configure-error-trackingSkill
- configure-lintingSkill
- configure-test-stackSkill
- configure-typescriptSkill
- create-moduleSkill
- optimize-performanceSkill
- scaffold-frontend-projectSkill
- set-up-authSkill
- set-up-design-systemSkill
- set-up-document-headSkill
- set-up-error-boundariesSkill
- set-up-feature-flagsSkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is 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.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 31
Never ask what the repo already answers.
- Distinctive: line 39
Stop at three rounds if the user answers "defaults": fill the rest from the radar and say so.
- Very specific: line 43
Write .claude/stack-profile.md (create .claude/).
≈ “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
Set Up Stack Profile
Writes .claude/stack-profile.md, the file every other skill reads first. Schema and precedence: stack-profile.md. Question wording and plugin map: wizard-reference.md.
1. Audit (change nothing)
cat .claude/stack-profile.md 2>/dev/null; cat ~/.claude/stack-profile.md 2>/dev/null
ls -a | head -50
ls pnpm-lock.yaml package-lock.json bun.lock bun.lockb yarn.lock uv.lock go.mod Cargo.toml pyproject.toml mise.toml .tool-versions .nvmrc justfile Makefile vercel.json supabase/config.toml drizzle.config.* sqlc.yaml biome.json* .golangci.y*ml 2>/dev/null
ls .github/workflows .gitlab-ci.yml 2>/dev/null; ls *.tf infra 2>/dev/null | head
grep -hE '"(react|vue|nuxt|next|hono|drizzle-orm|zod|@opentelemetry/[a-z-]+|@sentry/[a-z]+|@anthropic-ai/sdk|openai)"' package.json 2>/dev/nullApply the detection table in stack-profile.md to the results. Result: a draft profile with each key marked detected (with the file as evidence), inherited (from ~/.claude) or gap.
2. Decide
| State | Action |
| No project profile | Full run: steps 3–7. |
| Profile exists, nothing changed | Print "already in place" with the profile and stop. |
| Profile exists, detection differs or keys are missing | Delta run: show the diff (key: old → detected), ask only about changed and unknown keys. |
profile_version older than the schema | Migrate, show the diff, confirm. |
Never ask what the repo already answers. A detected value is shown for confirmation in the final summary, not asked as a question.
3. Ask only the gaps
Use AskUserQuestion, at most 4 questions per round, the recommended option first and labelled (Recommended). Round order: (1) language and framework gaps, (2) backend, database, hosting, (3) tooling (package manager, lint/format, task runner, tests, CI), (4) observability, AI. Skip a round when it has no gaps. Skip questions the answer of an earlier round made moot (no backend question for frontend-only static sites unless the user said otherwise).
Recommended options come from the Adopt rows in tech-radar.md. Before recommending a version-sensitive tool, check the live version per version-protocol.md. A deviation the user picks goes into the profile's notes with the reason.
Stop at three rounds if the user answers "defaults": fill the rest from the radar and say so.
4. Write the profile
- Write
.claude/stack-profile.md(create.claude/). Frontmatter exactly per the schema; notes below list deviations and why. - Keys the user marked
noneare written asnone; undecided keys are omitted. - Show the final file and the diff against the old one.
- Offer (do not do unasked) a user-wide default:
~/.claude/stack-profile.mdholding only taste keys (package_manager,runtime_manager,lint_format,tests,task_runner,ci,observability). Never copy project-specific keys (frontend,database,hosting) into it. - Suggest adding
.claude/stack-profile.mdto version control: it is team knowledge, not a secret.
5. Output: plugins to enable
Marketplace name is ai-development-skills; install id is <plugin>@ai-development-skills. Map profile → plugins with the table in wizard-reference.md. Always devcore. Print the exact commands:
/plugin marketplace add VelimirMueller/ai-development-skills
/plugin install devcore@ai-development-skills
/plugin install <other>@ai-development-skillsPlugins already enabled (/plugin lists them) are not repeated. Each plugin depends on devcore, so installing any of them installs it too.
6. Output: ordered skills to run
Print a numbered list, only skills that apply and whose result is not already in place. Order template and rules: wizard-reference.md. Typical start: validate-env → lint/format → TypeScript → tests → CI → security headers → logging/observability → feature skills.
7. Output: gaps no skill covers
For each profile choice with no matching skill in an installed or listed plugin (an unusual framework, a hosting provider not in the schema, other:<x> values), list it and offer extend-skillset. Do not run it unasked.
8. Verify
head -30 .claude/stack-profile.mdExpected: starts with ---, has profile_version: 1, parses as YAML, every value is in the schema's allowed list (or other:<name>). A second run of this skill prints "already in place".
Edge cases
- Monorepo: ask once whether to write one root profile or one per package; nearest profile wins.
- Conflicting evidence (two lockfiles): treat as a gap; ask which is canonical and suggest deleting the other.
- No
.git: still write the profile; mention that it is untracked.
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.

