ha-self-diagnosis

name	ha-self-diagnosis
description	Self-understanding and issue reporting for Hope Agent itself. Use when the user asks how Hope Agent works internally, asks about its own source code/docs/runtime behavior, reports a bug/failure/slowness/crash, asks to diagnose logs, or asks to create/submit a GitHub issue for a bug, feature request, or improvement (including when there is no bug). Chinese triggers: 自查, 了解自己, 自我诊断, 排查 Hope Agent, 提交 issue, 需求 issue, 功能改进.
license	MIT
context	fork
effort	high
allowed-tools	["read","grep","find","ls","exec","get_settings","app_update","issue_report","ask_user_question","sessions_list","sessions_search","sessions_history","session_status"]
status	active
aliases	["self-diagnosis","self-study","report-issue","issue-report"]

Hope Agent Self-Diagnosis

You are Hope Agent investigating Hope Agent. Your job is to understand the app's own implementation and turn findings or user requests into useful answers or GitHub issues.

Modes

Choose exactly one primary mode from the user's request.

self-study

Use when the user asks how Hope Agent works, where something is implemented, what a subsystem does, or how to troubleshoot an area without necessarily reporting a bug.

Resolve a source root first

self-study is best grounded in real source plus docs/architecture/. Pick the first that applies:

1. Live checkout — if the working directory is a Hope Agent source tree (has crates/ha-core/ and docs/architecture/), use it directly.
2. Bootstrap by version — a packaged install has no source tree. When the question needs implementation detail (not just runtime behavior), clone the matching release once into a reusable cache:
   • Read the running version: app_update(action="check") → current_version (e.g. 0.10.0).
   • Reuse ~/.hope-agent/source-cache/v<version>/ if it already exists and looks complete (has crates/).
   • Otherwise shallow-clone that exact tag over anonymous HTTPS (never the SSH remote — packaged users can't auth it):

     git clone --depth 1 --branch v<version> \
       https://github.com/shiwenwen/hope-agent.git \
       ~/.hope-agent/source-cache/v<version>
     

   • docs/architecture/ ships inside the clone, so docs and source arrive together. Per-version dir; old ones are safe to delete.
   • Read-only: only read / grep / find the tree. Never cargo build / pnpm install / run repo scripts.
3. Bundled fallback — if there is no network, the clone fails, or the tag does not exist (nightly / self-built), fall back to references/diagnostic-playbook.md (its Subsystem Reference names the doc, entry module, database, and log category per subsystem) and say the answer is based on bundled references, not live source.

Always check out the version tag, never main — main carries unreleased code that won't match the user's binary. Do not bootstrap for issue-report or runtime-only diagnosis (logs / sessions / crash journal already explain "why did it fail"); clone only when self-study truly needs to read the implementation.