jgrep Agent

jgrep-agent is a skills-only Agent Plugin for Codex. It teaches coding agents to use jgrep's compact --ai mode to find code locations, then inspect only the source ranges that matter. It adds no network service, background process, or model of its own.

Use the current jgrep installer before enabling the plugin so that jgrep is available on the coding environment's PATH. See the installation and agent integration guide.

Install in Codex

This command registers the repository marketplace, installs the plugin, and does not require cloning the repository:

codex plugin marketplace add xxvw/jgrep --ref main --sparse .agents/plugins --sparse plugins/jgrep-agent && codex plugin add jgrep-agent@jgrep

Start a new Codex session after installation so the skill is available. In the ChatGPT desktop app, select the jgrep marketplace in the Plugins Directory and install jgrep-agent.

What the skill changes

For a conceptual search, it prefers semantic jgrep --ai:

jgrep --ai --ai-max-results 25 -r --include '*.rs' \
  'where authentication failures are handled' src/

For a known literal or pattern, it prefers -F or -E. Those lexical modes do not download or load the local model:

jgrep --ai --ai-max-results 25 -r -F 'validate_session' src/
jgrep --ai --ai-max-results 25 -r -E 'Auth(Error|Failure)' src/

The output is one locator per selected line:

src/auth/session.rs:42
-:7

Parse the final colon before the decimal line number, since Windows paths can contain a drive-letter colon. The default cap is 50 locations. If jgrep reports that the cap was reached on stderr, the result is incomplete; narrow the search before raising --ai-max-results. The skill treats exit status 1 as a normal no-result outcome and falls back to rg when jgrep is not available.

--ai is deliberately incompatible with output modes that hide or expand locations, including -c, -l, -L, -q, -m, context flags, --score, and --color=always. The full rules are in skills/jgrep-agent/SKILL.md.

Read this plugin in your language

License

This plugin is part of jgrep and is licensed under GPL-3.0-or-later.