agentleFS
badmoonrbadmoonr/loopilot/skills/loopilot-doctor/SKILL.md

loopilot-doctor

Skill1 starsChanged 27 days agoApache-2.0View on GitHub

Run read-only diagnostics for a macOS LooPilot installation and generate a fail-closed redacted support bundle. Use to inspect installed release parity, configuration, providers, Chrome/CDP, Dreamina paths, locks, queues, SQLite, manifests, outbox, launchd, workbench health, and recent run state without repairing or restarting anything.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 2 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run diagnostics without changing configuration, locks, queues, services, sessions, or business state.

    line 11
  2. Very specific:

    Run scripts/doctor.py run with the installed deployer script and the exact install/config/state/run roots.

    line 15
  3. Very specific:

    If support evidence is needed, write the local doctor report, then run support-bundle with every user-owned root passed as --redact-root.

    line 20
  4. Distinctive:

    Never clear a lock, rewrite SQLite, log in, resolve a secret value for display, restart launchd, start Chrome, refresh cookies, or resume a …

    line 45
  5. Very specific:

    Use doctor-run-config.json as the handoff checklist for customer-specific roots and account keys.

    line 47

≈ “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

LooPilot Doctor

正式独立运行时遵循任务方案、工作台登记与主动定位:主 Agent 登记真实任务并打开当前聊天对应详情;子阶段沿原任务,后台仅更新数据。此约定不扩大本 Skill 的业务权限。

Run diagnostics without changing configuration, locks, queues, services, sessions, or business state. Read support-bundle-contract.md before exporting diagnostics.

Workflow

  1. Run scripts/doctor.py run with the installed deployer script and the exact install/config/state/run roots. Only standard <run>/reports/report_manifest.json files are validated as report-manifest-v1; same-named analytical outputs under outputs/ are outside this contract.
  2. Review blocked and warning checks. A warning is not a repair instruction. Ark 媒体数据面检查只读验证 Provider Profile v3、Keychain availability、Snapshot hash、账号 scope 与 release lane;不得执行 arkcli auth status/profile show/resources list/models get,也不得把 SSO 登录态作为通过条件。Codex harness 的显式控制面诊断继续单独显示,不能替代数据面结论。 LP-082 启用时还要验证 sidecar active/release hash、Python/adapter smoke、固定服务账号 acquisition profile 和 rollout;cookies 仅 lstat owner/mode/regular/symlink,不打开、不解析、不计算 hash。 混剪客户环境额外只读检查 Apple 芯片、FFmpeg/ffprobe、11-Skill release 闭包、loopilot-account-v3/混剪 profile、账号素材库、TTS Keychain 引用与账号 scope、以及 workbench loopback;不读取 Keychain 值、不创建素材目录、不启动服务。
  3. If support evidence is needed, write the local doctor report, then run support-bundle with every user-owned root passed as --redact-root.
  4. Re-scan the generated bundle. If a secret, signed URL, user-home path, browser profile, binary, or real media is detected, deliver nothing.
  5. Hand any approved repair to $loopilot-deployer with a new scoped authorization file.

Commands

python3 scripts/doctor.py run \
  --deployer-script <install-root>/loopilot-deployer/scripts/deployer.py \
  --install-root <install-root> \
  --config-root <config-root> \
  --deployment-state-root <deployment-state-root> \
  --state-root <merchant-state-root> \
  --run-root <merchant-run-root> \
  --output <doctor-report.json>

python3 scripts/doctor.py support-bundle \
  --doctor-report <doctor-report.json> \
  --output-dir <empty-output-dir> \
  --redact-root <install-root> \
  --redact-root <config-root> \
  --redact-root <state-root> \
  --redact-root <run-root>

Never clear a lock, rewrite SQLite, log in, resolve a secret value for display, restart launchd, start Chrome, refresh cookies, or resume a task. Report the finding and stop.

Use doctor-run-config.json as the handoff checklist for customer-specific roots and account keys. Do not store credentials in this file.

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.