omp packages

@diegopetrucci/pi-repo-scout0.1.0

A standalone pi extension that adds a read-only repo_scout subagent tool that scans a repository and reports stack, conventions, commands, and hotspots in an isolated child agent.

diegopetrucci Published 7 days ago MIT ~358 downloads/mo

Untested on omp

This is a pi package. No one has reported whether it works on omp yet. If you try it, add your result below.

Did it work for you?

Result

One report per person. A new report replaces your old one.

Reports 0

No reports yet.

Readme

repo-scout

Adds a repo_scout subagent tool that adapts The Last Harness repo-scout prompt into a standalone pi extension.

Use it to quickly orient in an unfamiliar repository before planning or implementing. The subagent uses read-only built-in tools plus any configured MCP tools, detects the stack from evidence, and reports conventions, validation commands, hotspots, and open questions with evidence paths. Blockers are reported in the final result; the subagent cannot ask follow-up questions.

Install

Standalone npm package

pi install npm:@diegopetrucci/pi-repo-scout

Collection package

pi install npm:@diegopetrucci/pi-extensions

GitHub package

pi install git:github.com/diegopetrucci/pi-extensions

Then reload pi:

/reload

repo_scout tool behavior

The tool accepts:

  • task — optional focus for the scan (for example the area the caller is about to work on); when omitted, a general orientation report is produced
  • cwd — optional working directory (repository) to scan; defaults to the current session directory

The subagent runs in-process in an isolated child session with:

  • the current session model (no automatic model selection and no persisted /config commands);
  • medium thinking, clamped to the model's supported levels;
  • read-only built-in tools: read, grep, find, ls, and guarded bash;
  • configured MCP tools (run without confirmation in the subagent; may have side effects; use for evidence gathering only; requires host Pi >=1.0.4; older hosts degrade to no MCP, no error); the global mcp.json always applies; a project .pi/mcp.json is loaded only when the host session has already trusted the project;
  • a local-checkout path guard for file inspection;
  • a bash guard that allows only direct read-only git, gh, or pwd invocations.

The output is a single markdown report: detected stack, conventions, commands, project hotspots, do/don't patterns, and open questions. The tool appends a run-details suffix with the model, effective thinking level, turn count, tool calls, duration, working directory, and tools used.

Read-only guarantees

  • The subagent is explicitly instructed not to modify anything.
  • Runtime guards block write/edit tools, shell control operators, pipelines, redirection, path traversal outside the checkout, mutating git/gh commands, npm/publish commands, and other filesystem mutation.
  • Runtime guards apply to built-in tools only; configured MCP tools pass the guard, run without confirmation, and are limited only by prompt guidance.
  • Built-in file-inspection tools are preferred over shell commands for local files.

Example

Use repo_scout to orient me in this repository before I add a new command:

{
  "task": "Where do CLI commands live and how are they tested?"
}

Limitations

  • The subagent has an 8-turn and 8-minute budget and a default 30-second timeout per bash call (applied when the model does not set one); tool use is disabled on the final turn so it can answer.
  • The report samples representative files only; it is an orientation aid, not an exhaustive review.
  • gh-based inspection is available only when GitHub CLI is installed and authenticated; the tool works on local repositories without it.