add-stack-context

Audit project stack technologies and install official vendor agent context.

Updated Sep 17, 2026
One-click install
npx skills add https://github.com/hkw2028/toy-project --skill add-stack-context-hkw2028
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: add-stack-context
Source: https://github.com/hkw2028/toy-project/tree/main/.agents/skills/add-stack-context
Command: npx skills add https://github.com/hkw2028/toy-project --skill add-stack-context-hkw2028

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Coding agents often work without access to official, version-accurate guidance for the frameworks, libraries, and services a project uses, leading to outdated or hallucinated patterns. This Skill audits the project's declared stack and ensures every technology has current, vendor-controlled agent context installed. ## Core Features & Use Cases - Stack Inventory: Builds a checklist of direct dependencies from manifests and config files, using lockfiles only to confirm installed versions without expanding into transitive packages. - Official Context Discovery: Uses find-skills or the Skills CLI to locate vendor-controlled skills, verifying ownership against official documentation rather than search rank or install counts. - Live Guidance Management: Adds bounded managed instructions to AGENTS.md or CLAUDE.md that point agents to retrieve current vendor sources like llms.txt at task time, while preserving user-authored content. - Use Case: When adopting a new project with Next.js, Prisma, and Playwright in package.json, run this Skill to install each vendor's official agent context and record gaps where no official source exists. ## Quick Start Ask the agent to audit this project's stack and install official agent context for every direct dependency.

Frequently Asked Questions about add-stack-context

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I set up official agent context for my project's tech stack?

Run the stack context audit, which inventories direct dependencies from manifests and config files, then discovers and installs vendor-controlled skills or official context like codemods and llms.txt references for each technology.

How does the audit decide which skills are official versus community-made?

It verifies candidate ownership against the technology vendor's official documentation or organization, not search rank or install counts. Community skills are only installed after explicit user approval and never substitute for missing official context.

What happens if a library has no official agent context available?

The gap is recorded as unavailable from an official source, and the audit continues with the remaining stack items. A community candidate may be listed separately but is never treated as a substitute for official context.

Does it work when the find-skills helper is not installed?

Yes. The missing helper is treated as a fallback condition, and the audit searches the current Skills ecosystem through the Skills CLI or an equivalent source, then continues to non-skill vendor context like installers and codemods.

Will it overwrite my existing AGENTS.md or CLAUDE.md rules?

No. User-authored content and existing links between instruction files are preserved. It only adds or updates a bounded managed instruction, and vendor-managed blocks are updated solely through the vendor's documented method.

Should llms.txt content be copied into the repository?

No. A vendor llms.txt is treated as a live remote source. The audit adds an instruction telling later agents to retrieve the current document when relevant work begins and reconcile it with the installed version.