One-click install
npx skills add https://github.com/ronaknnathani/relay --skill explore-ronaknnathani
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: explore
Source: https://github.com/ronaknnathani/relay/tree/main/skills/explore
Command: npx skills add https://github.com/ronaknnathani/relay --skill explore-ronaknnathani

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Agents and developers often act on a codebase without verified context, leading to wrong assumptions about entry points, data flow, and module boundaries. This Skill builds a trustworthy, citation-backed map of a codebase or a single feature slice so downstream planning, implementation, and review work from facts instead of guesses. ## Core Features & Use Cases - Stack Detection: Reads dependency manifests (go.mod, package.json, pyproject.toml, Cargo.toml, and more) to confirm languages, frameworks, and versions before making any claim. - End-to-End Code-Flow Tracing: Follows call chains from entry points through every abstraction layer down to where data is persisted or leaves the process, recording each hop as file:line references. - Architecture Analysis: Documents module boundaries, ownership, dependency direction, seams, and deep versus shallow modules using a shared vocabulary downstream skills can reuse. - Use Case: Before planning a change to an API's request validation, a planning skill calls this Skill to map the handler entry point, the service seam, and the persistence boundary, receiving a cited report with an essential-files list. ## Quick Start Ask the agent to explore how authentication requests flow through this repository and report the entry points, seams, and essential files with file:line citations.

Frequently Asked Questions about explore

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

FAQPage Schema
How do I map a codebase before making changes?▼

Start by reading dependency manifests like go.mod or package.json to confirm the stack, then trace each entry point through every layer to where data is persisted. Record findings as file:line citations and close with a short essential-files list.

How to trace code flow across abstraction layers?▼

Grep for the feature's routes, types, and config keys to find entry points, then follow the call chain through each interface down to the database, queue, or network boundary. Note every seam crossed and whether the module behind it is deep or shallow.

Can this Skill run tests or builds to verify behavior?▼

No, it is strictly read-only and never runs build, test, or codegen commands because they can mutate state through caches, installs, or network calls. It identifies and cites the relevant commands so a downstream skill can run them instead.

What happens when a code claim cannot be verified?▼

Unconfirmed claims are explicitly tagged as unverified along with what evidence would confirm them. The Skill never presents an inference as fact, and every other claim must carry a file:line reference.

When should I explore a slice instead of the whole repository?▼

Scope the exploration to the topic the caller named whenever the request targets one feature or area. Mapping the entire repository when asked about a slice wastes effort and dilutes the report with irrelevant files.