Hs init
Create or refresh a concise project map for AI-assisted development. Use when a project needs an AI entry point, architecture overview, coding conventions, product/domain context, or a synchronized set of project orientation documents.From its SKILL.md
npx -y skills add danielleit241/hs-skills --skill hs-initAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 20 days oldThe repository was created 20 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
3.4 KB, 586 tokens by cl100k_base, as published. Nobody here has run it
Init - Project Map
Create a compact, evidence-based map of an existing project so an AI agent can understand context and navigate the codebase quickly.
Principle
Produce a map, not a rulebook. Describe the current project as observed; do not invent mandatory policies, prescribe implementation details, or turn preferences into hard rules.
Workflow
- Inspect the repository structure, entry points, manifests, configuration, tests, deployment files, and existing documentation.
- Identify the project purpose, runtime boundaries, major components, data or request flow, development conventions, and important navigation paths.
- Read existing
CLAUDE.md,AGENTS.md, and target documents before changing them. Preserve useful project-specific context. - Read
references/project-map-best-practices.mdbefore drafting. Use repository evidence to describe the current state and its research-backed include/exclude guidance to keep each document focused. Research stack-specific gaps only from authoritative sources; add a source to a generated document only when it supports useful project-specific guidance. - Create or update only the orientation documents that already exist, the user explicitly requests, or repository evidence shows are needed:
CLAUDE.mdorAGENTS.md: an AI entry point with repository orientation and links to deeper documentation. Update both only when both exist or both audiences need distinct guidance.system-architecture.md: current architecture, boundaries, integrations, and important flows.code-standards.md: observed conventions for naming, testing, errors, tooling, and validation.project-overview-prd.md: product goal, users, domain concepts, workflows, and business context.
- Link documents from applicable entry points where useful, remove stale claims, and keep each document short enough to scan.
- Report unknowns explicitly instead of guessing. Do not modify application code unless the user separately requests it.
Update Policy
- Refresh the map when architecture, product scope, tooling, or conventions change significantly.
- Prefer small targeted updates over rewriting accurate sections.
- Separate observed facts, inferred relationships, and unresolved questions.
- Apply the reference's research-backed inclusion and exclusion guidance. Do not add generic best-practice sections or unverified preferences to the project map.
- Treat these documents as orientation material; implementation plans and repository-specific rules remain authoritative elsewhere.
Quality Check
- Confirm every claim is supported by current repository evidence or clearly marked as unresolved.
- Confirm paths, commands, component names, and integrations still exist.
- Confirm applicable entry points guide navigation without duplicating large reference documents.
- Confirm each created or updated document is justified by repository evidence or an explicit user request, remains concise, and has a distinct purpose.
Read references/project-map-best-practices.md before drafting and references/document-templates.md for the document skeletons and evidence checklist.