# Feature Marker

> Gemini-ready install target for the feature-marker run-through skill workflow. Use from a Gemini prompt; the npm package is only the installer.

- Skill: `majiayu000/feature-marker` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add majiayu000/feature-marker`
- Raw SKILL.md: https://api.skillmd.com/api/skills/majiayu000/feature-marker/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: majiayu000 (https://skillmd.com/u/majiayu000)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/majiayu000/feature-marker

---


# feature-marker for Gemini

Use this skill from inside Gemini prompts:

```text
Use feature-marker to implement billing-observability.
```

Run the complete flow by default:

- Create or reuse `tasks/{slug}/prd.md`, `tasks/{slug}/techspec.md`, and
  `tasks/{slug}/tasks.md`.
- Use the bundled templates in `templates/` for PRD, TechSpec, and Tasks; fill
  `{slug}` and `{feature_title}` and leave no unresolved placeholders.
- Create or require a feature branch before implementation.
- Use a git worktree only when the checkout is dirty or the user asks.
- Run an implementation grill pass before coding to find artifact, task, test,
  risk, and handoff gaps.
- Resolve grill findings in the artifacts, asking the user only when a finding
  changes scope or requires a product decision.
- Stop only for true ambiguity, unrelated dirty work, or blocked verification.
- Run verification, commit locally, and print exact push/PR handoff commands.
- Do not push, open a PR, create checkpoint JSON, or use an interactive menu.

`spec-driven` and `ralph-loop` are out of scope for this skill-first v1 unless
they are rebuilt as explicit skill instructions.