Skip to content
Back to skills

Init

ASecurity

Interactive setup of .slimspec/config.yaml in the session.

  • 3 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added November 7, 2026
ai-agentsgogit

Works with

  • claude code

Security analysis

A100/100

Scanned November 11, 2026

npx -y skills add yushman/slimspec --skill init --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Init?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Init
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/yushman-init/badge)](https://www.skillsdirectory.com/skills/yushman-init)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: init
description: "Interactive setup of .slimspec/config.yaml in the session."
disable-model-invocation: true
---

The config is optional: every key has a default. Ask with `AskUserQuestion`; each option gets a one-line description of its meaning.

1. If `.slimspec/config.yaml` exists, show it; its current values are the recommended (first) options below. Otherwise the defaults are.
2. Detect the test command by reading files: `package.json` with a `"test"` script -> `npm test`; `Makefile` with `test:` -> `make test`; `gradlew` -> `./gradlew test`; `Cargo.toml` -> `cargo test`; `go.mod` -> `go test ./...`; `pyproject.toml` or `pytest.ini` -> `pytest`. First match wins.
3. One call, 4 questions:
   - `mode`: fast | medium | full.
   - `spec.format`: claude-md | readme | central | native (4 options max; name `agents-md` and `none` in the question text, reachable via Other).
   - `test`: auto | the detected command, if any; an explicit command via Other.
   - `commit`: none | lane | wave.
4. One call:
   - `apply.agents` (max parallel implementers): 3 | 1 | 2 | 5; another integer >= 1 via Other.
   - `review.depth`: auto | low | medium | high.
   - Only when `spec.format` is not `none`: `spec.unit` (auto | a dir glob via Other, e.g. `src/*/`) and `spec.adr` (no | yes: also keep decisions in `docs/adr/`).
5. One call:
   - `changes.closed` (closed change folders): delete | archive (moved to `.slimspec/archive/<date>-<slug>/`).
   - `changes.gitignore`: yes (add change folders to `.gitignore`) | no (tracked; change files go into lane or wave commits).
6. Models, two calls: scanner, debugger, planner, then implementer, reviewer, grounder. Recommended (first) option per role: its existing `models.<role>`, else its D16 tier (heavy: debugger, planner, reviewer, grounder; light: scanner, implementer).
   - Claude Code: opus | sonnet | haiku; the default tier is the role's agent frontmatter model. Write only roles whose answer differs from it.
   - Other hosts: the options and how to list them come from the host mapping's Models section. Write every answer to `models`, except the session model, which writes no key, unless the mapping says otherwise.
7. Show the resulting YAML in the exact layout of the config block in [protocol.md](../slimspec/protocol.md) (same keys, aligned `#` comments; quote `spec.unit` and `test` unless `auto`), then ask: write | change. On change, re-ask only the keys the user names.
8. On write, create `.slimspec/` and write `.slimspec/config.yaml`. Never touch `.gitignore`. Reply with one line: the written path; models can be edited in the file.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.