Find repository context for coding agents with Jevgrep
Ask a natural-language codebase question and return ranked files, source excerpts, and reading leads so a coding agent can start work with grounded context.
npx skills add agentskillexchange/skills --skill find-repository-context-for-coding-agents-with-jevgrep
npm install --global @dzhng/jevgrep@latest, run jg auth to choose and save a supported provider credential, then run jg skill or npx skills add dzhng/jevgrep --skill jevgrep from the project where the coding agent works.Use Jevgrep when a coding agent needs to understand an unfamiliar repository behavior but does not yet know the right files or symbols to inspect. The operator installs the `jg` CLI and project skill, authenticates a supported provider, then asks targeted repository questions such as where telemetry is recorded, where auth is enforced, or which tests cover retry behavior. Jevgrep returns a compact stdout packet with a summary, ranked file list, verbatim source excerpts, and line-referenced reading leads. Keep the scope to discovery and context retrieval: the agent still reads the cited files, edits code, and runs tests separately.
What this skill actually does
Inputs and prerequisites: Node.js 22+, npm, jg CLI, supported Jevgrep provider credentials.
Setup notes: Install with npm install –global @dzhng/jevgrep@latest , run jg auth to choose and save a supported provider credential, then run jg skill or npx skills add dzhng/jevgrep –skill jevgrep from the project where the coding agent works.
Source and verification boundary: use https://github.com/dzhng/jevgrep as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.