agentleFS
NVIDIAKnown publisherNVIDIA/nemoclaw-community/examples/recipes/nvidia/agentic-ai-learning-path/agents/hermes/skills/nemoclaw-autoheal/SKILL.md

nemoclaw-autoheal

Skill142 starsChanged 3 months agoApache-2.0View on GitHub

Guide users through Hermes availability checks and the optional host-side auto-heal setup without crossing the sandbox-to-host boundary.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    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. Very specific:

    Use this skill when a user says the Slack bot is not responding, asks whether Hermes is healthy, reports inference 503/504 failures, asks …

    line 13
  2. Very specific:

    Call nemoclaw_status first.

    line 26
  3. Distinctive:

    Run these commands from the cloned example directory, not inside the sandbox:

    line 39
  4. Distinctive:

    If auto-heal is not installed yet, first run:

    line 46
  5. Very specific:

    If the user runs an inference endpoint through http://host.openshell.internal:18080, remind them to set NEMOCLAW_HOST_TLS_PROXY_UPSTREAM in …

    line 61
  6. Distinctive:

    Start with the current status or likely failure class, then give only the next one or two host commands.

    line 69

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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
<!-- SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. SPDX-License-Identifier: Apache-2.0 -->

nemoclaw-autoheal

Use this skill when a user says the Slack bot is not responding, asks whether Hermes is healthy, reports inference 503/504 failures, asks about Outlook connectivity, or wants to install the optional auto-heal services.

Boundary

You run inside the OpenShell sandbox. Do not run systemctl, recreate the sandbox, inspect credentials, or try to install host services. Those actions belong to the host operator. You may call nemoclaw_status to inspect the local Hermes gateway, then give the operator the exact commands below.

Diagnose

  1. Call nemoclaw_status first. If the gateway is stopped, say that the sandbox gateway is unhealthy.
  2. For an inference error, explain that 503, 504, or a timeout can be an upstream service or host-proxy failure.
  3. For a Slack failure, distinguish an unavailable gateway from an allowlist, Socket Mode, or Slack API issue. Do not claim the bot received a message unless the user supplied evidence.
  4. For Outlook, explain that a Graph connection closure may be sandbox egress or a temporary proxy failure; it is not proof that the mailbox settings are wrong.

Tell the host operator

Run these commands from the cloned example directory, not inside the sandbox:

bash scripts/autoheal/sanity-check.sh
bash scripts/autoheal/sanity-check.sh --repair

If auto-heal is not installed yet, first run:

bash scripts/autoheal/install.sh --check
bash scripts/autoheal/install.sh

For status and logs:

systemctl --user status nemoclaw-hermes-gateway-forward.service
systemctl --user list-timers 'nemoclaw-*'
journalctl --user -u nemoclaw-hermes-watchdog.service -n 80 --no-pager

If the user runs an inference endpoint through http://host.openshell.internal:18080, remind them to set NEMOCLAW_HOST_TLS_PROXY_UPSTREAM in the host .env before installing the auto-heal services. The value is the HTTPS origin, for example https://inference-api.nvidia.com.

Response style

Start with the current status or likely failure class, then give only the next one or two host commands. Do not print or request Slack, Outlook, GitHub, or inference credentials.

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.