agentleFS
asta-nguyenasta-nguyen/agent-devkit/skills/setup-openez/SKILL.md

setup-openez

Skill4 starsChanged 2 months agoMITView on GitHub

Use when the user asks to set up OpenEZ, or when semantic code search is needed and its workspace index is missing or stale.

  • Installs packages

1 other project uses a version of this, 1 of them edited.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

17 rules. About 0 are common, 6 are distinctive, and 11 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 11

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Resume after client wiring: If a previous session completed client wiring, skip steps 1–4 and go to step 5.

    line 24
  2. Very specific:

    Initialize the workspace index: Ask the user before running openez init (it creates a .openez/ directory of derived index data) and wait …

    line 27
  3. Distinctive:

    Index the repository when needed: Ask the user first and wait for approval.

    line 38
  4. Distinctive:

    Use it for initial setup or when the index is missing or stale.

    line 45
  5. Very specific:

    If AGENTS.md exists and does not mention OpenEZ, present the exact block in ## Code intelligence block below and wait for the user's …

    line 52
  6. Very specific:

    If the user declines, leave AGENTS.md unchanged and report that the guidance was not added.

    line 56
  7. Very specific:

    Read openez setup --help to find supported clients.

    line 62
  8. Distinctive:

    If the user's client is not listed, consult that client's official MCP configuration documentation (for example, Cursor).

    line 69
  9. Very specific:

    If the user approves and wiring is applied, stop and tell the user to restart the agent.

    line 80
  10. Very specific:

    If client wiring was approved and run, verify in a session after restart.

    line 88
  11. Very specific:

    When OpenEZ MCP tools are available in this session, call code_query with a simple query (e.g., "entry point" or "main function").

    line 89
  12. Distinctive:

    Check that the MCP server is running.

    line 94
  13. Very specific:

    If still failing, fall back to direct file reads and rg.

    line 96
  14. Very specific:

    If OpenEZ MCP tools are unavailable in this session, do not call code_query.

    line 98
  15. Distinctive:

    After significant code changes (new files, renamed modules, large diffs)

    line 127
  16. Very specific:

    When code_query returns stale or missing results

    line 128
  17. Very specific:

    Before document-wiki or read-codebase-context on a fresh checkout when OpenEZ is in use and its index is missing or stale

    line 129

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Setup OpenEZ

OpenEZ provides MCP tools for semantic code queries, graph traversal, and memory recall. This skill bootstraps it for a repository.

Prerequisites

  1. Bun 1.1+ — check with bun --version. If missing, ask the user to authorize installation before continuing. Do not install silently.
  2. OpenEZ CLI — check with openez --version. If missing, explain that a global CLI install is needed and ask the user to authorize it before running:

    npm install -g @openez-graph/cli

Process

Resume after client wiring: If a previous session completed client wiring, skip steps 1–4 and go to step 5. Re-index only if the index is missing or stale, and only after the user approves it (step 2).

  1. Initialize the workspace index. Ask the user before running openez init (it creates a .openez/ directory of derived index data) and wait for approval. Ensure .gitignore contains .openez/; append it only when missing and preserve existing content. Never untrack an existing .openez/ path automatically; report it instead. If .openez/ does not exist and the user approved, run:

    openez init .
  2. Index the repository when needed. Ask the user first and wait for approval; indexing a large repository can take a while. Then run:

    openez index .

    Use it for initial setup or when the index is missing or stale. Report progress and check the command result: a nonzero exit means indexing failed; do not report the index as created. A successful CLI run confirms that command completed, not that the MCP query works.

  3. Record in AGENTS.md:

    If AGENTS.md exists and does not mention OpenEZ, present the exact block in ## Code intelligence block below and wait for the user's approval before adding it.

    If the user declines, leave AGENTS.md unchanged and report that the guidance was not added. If AGENTS.md does not exist, skip this step; setup-codebase owns its creation.

  4. Wire the agent client (optional; ask the user first):

    Read openez setup --help to find supported clients. For a supported client, get approval before running:

    openez setup <supported-client>

    If the user's client is not listed, consult that client's official MCP configuration documentation (for example, Cursor). Do not invent a configuration snippet. Present any documented configuration change for approval, then apply the same restart-and-verify procedure below. If no supported configuration can be established, report that limitation and continue to step 5 without wiring.

    Wiring configures the agent's MCP client to use OpenEZ. The user must restart their agent after this step for MCP tools to load. Do not run this without approval; it changes agent-specific configuration.

    If the user approves and wiring is applied, stop and tell the user to restart the agent. Do not continue to step 5 in this session. Tell the user to invoke setup-openez again after restarting so the new session can verify the connection; an agent cannot restart its own session and continue here. If wiring is declined or skipped, continue to step 5 in this session.

  5. Verify the MCP connection:

    If client wiring was approved and run, verify in a session after restart. When OpenEZ MCP tools are available in this session, call code_query with a simple query (e.g., "entry point" or "main function"). If the query returns results, the MCP server is connected and the index is working. If it fails or returns nothing:

    • Check that the MCP server is running.
    • Re-index only if the index is missing or stale.
    • If still failing, fall back to direct file reads and rg.

    If OpenEZ MCP tools are unavailable in this session, do not call code_query. If wiring was declined or skipped, report that the client connection remains unconfigured. If wiring was approved and run, report that the MCP tools did not load after restart. In either case, report the MCP index query as index unverified. If the CLI indexing command succeeded, report that separately as CLI indexing completed; if it failed or was not run, report the index as failed or unverified. Continue with direct file reads and rg.

Code intelligence block

This is the only definition of the AGENTS.md section; setup-codebase uses it too.

## Code intelligence

Locate concepts with OpenEZ `code_query`, approximate filenames with FFF
fuzzy file find, identifiers or literals with FFF grep (fallback: `rg`),
and regex with `rg`. Expand callers and callees with OpenEZ `code_context`
(1–2 hops); confirm string, route, config-key, and case variants with FFF
multi-pattern grep (fallback: `rg`). Use OpenEZ `code_outline` before
reading a large file. Search results are navigation; read current source as
evidence. Query OpenEZ directly with the repository path; if unavailable,
unindexed, or irrelevant, continue with FFF or `rg` and direct reads.

When to re-index

  • After significant code changes (new files, renamed modules, large diffs)
  • When code_query returns stale or missing results
  • Before document-wiki or read-codebase-context on a fresh checkout when OpenEZ is in use and its index is missing or stale
openez index .

Re-indexing is incremental and safe to run repeatedly, but only re-index when one of these conditions applies.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.