Explain
Use this skill to explain the project at the current working directory to the user. This is different from onboarding: the goal is not only to orient the agent, but to produce a clear user-facing explanation.
Scope
Explain:
- what the project appears to do
- who or what it is for
- main technologies and frameworks
- important files and directories
- how the code is organized
- how to run, build, and test it when discoverable
- noteworthy constraints, risks, or unclear areas
Workflow
- Inspect project instructions first.
- Read local
AGENTS.md, .codex/, .agents/, README files, and central docs when present.
- Respect project-specific instructions over global assumptions.
- Identify project type and stack.
- Check common manifest/config files such as
package.json, pyproject.toml, Cargo.toml, go.mod, platformio.ini, CMakeLists.txt, compose.yaml, and CI workflows.
- Map structure without over-reading.
- Prefer
rg --files and shallow directory listings.
- Avoid dependency, generated, cache, and build output directories.
- Read representative files.
- Inspect entry points, main modules, routes, commands, hardware targets, or app screens depending on the project.
- Read only what is needed to explain the project accurately.
- Produce a user-facing explanation.
- Keep it concise and concrete.
- Separate confirmed facts from inferences.
- Mention unknowns only when they matter.
- Do not edit files unless the user explicitly asks for documentation changes.
Output
Use this structure when helpful:
- Overview
- Main Stack
- Directory Map
- How It Works
- Run / Build / Test
- Important Files
- Notes and Unknowns
Prefer short sections over long exhaustive analysis. If the user asks for a README-style document, propose a target file and wait for explicit edit approval before creating or modifying files.
1---2name: explain-23description: Use when the user asks to explain the project in the current working directory, summarize what this repository does, describe its structure, or provide a user-facing project overview. This skill performs read-only inspection and returns a concise explanation for the user.4---56# Explain78Use this skill to explain the project at the current working directory to the user. This is different from onboarding: the goal is not only to orient the agent, but to produce a clear user-facing explanation.910## Scope1112Explain:1314- what the project appears to do15- who or what it is for16- main technologies and frameworks17- important files and directories18- how the code is organized19- how to run, build, and test it when discoverable20- noteworthy constraints, risks, or unclear areas2122## Workflow23241. Inspect project instructions first.25 - Read local `AGENTS.md`, `.codex/`, `.agents/`, README files, and central docs when present.26 - Respect project-specific instructions over global assumptions.272. Identify project type and stack.28 - Check common manifest/config files such as `package.json`, `pyproject.toml`, `Cargo.toml`, `go.mod`, `platformio.ini`, `CMakeLists.txt`, `compose.yaml`, and CI workflows.293. Map structure without over-reading.30 - Prefer `rg --files` and shallow directory listings.31 - Avoid dependency, generated, cache, and build output directories.324. Read representative files.33 - Inspect entry points, main modules, routes, commands, hardware targets, or app screens depending on the project.34 - Read only what is needed to explain the project accurately.355. Produce a user-facing explanation.36 - Keep it concise and concrete.37 - Separate confirmed facts from inferences.38 - Mention unknowns only when they matter.39 - Do not edit files unless the user explicitly asks for documentation changes.4041## Output4243Use this structure when helpful:4445- Overview46- Main Stack47- Directory Map48- How It Works49- Run / Build / Test50- Important Files51- Notes and Unknowns5253Prefer short sections over long exhaustive analysis. If the user asks for a README-style document, propose a target file and wait for explicit edit approval before creating or modifying files.
kotek-7