Codebase Context
Purpose
This is a background knowledge skill. It's not invoked directly by users — it's referenced by other skills and agents when they need workspace context.
What to Gather
When activated, build a context summary covering:
Project Identity
- •Repository name and purpose (from README)
- •Technology stack (from package manifests and lock files)
- •Framework versions (from lock files, not manifests)
- •Build system and scripts
Structure
- •Top-level directory layout and purpose of each directory
- •Entry points (main files, config, routes)
- •Test directory structure and framework
Conventions
- •Naming patterns (files, classes, methods, variables)
- •Architectural patterns in use (MVC, services, etc.)
- •Error handling patterns
- •Logging patterns
Configuration
- •Environment variables in use
- •Config file locations
- •CI/CD pipeline structure
Accumulated Knowledge
- •Read
.github/agent-context.mdfor previously discovered patterns - •Read
docs/solutions/index for documented learnings
Output Format
markdown
## Workspace Context ### Stack [Language] [Framework] [Version] — [Build tool] ### Key Paths - `path/` — [purpose] ### Conventions - [Convention 1] - [Convention 2] ### Accumulated Knowledge [Summary from agent-context.md and docs/solutions/]
Guardrails
- •Read-only. Do not modify any files.
- •Focus on facts discoverable from the codebase, not assumptions.
- •Keep the summary under 200 lines — this will be injected into agent context.