OpenClaw Ollama Runtime KB

Scope

• In-scope: quick lookup/start-stop-status-log/pull model/debug for OpenClaw Ollama Runtime on Android.
• Out-of-scope: patch runtime outside SOP KB or change provider policy that is not required.

SOP

1. Determine whether the request is within the scope of the skill or not.
2. Execute the skill's main workflow/instructions in a safe order.
3. Check the results, state validation and next steps.

Indexed Reference Map

• @[SKILL.md#L1-L70] Frontmatter + target + skill range.
• @[SKILL.md#L25-L70] Usage procedures and guardrails.
• @[references/usage-cheatsheet.md#L1-L64] Common command syntax + operating workflow.
• @[references/troubleshooting.md#L1-L87] Diagnose errors by symptoms (401/404/unhealthy/context).
• @[references/model-management.md#L1-L40] Pull/list/test model + naming conventions.
• @[scripts/diag.sh#L1-L41] Script to quickly check runtime health.

KB Mode Notes

• Reading order:
  1. references/INDEX.md
  2. references/usage-cheatsheet.md
  3. references/troubleshooting.md
  4. references/model-management.md
• Coverage Matrix:
  • day-to-day commands -> usage-cheatsheet.md
  • symptom-based diagnosis -> troubleshooting.md
  • pull/list/test model flow -> model-management.md
• Portability note:
  • This is the operating KB for OpenClaw Ollama Runtime on Android; Do not use to justify patch runtime outside the lookup/diagnosis scope.

Target

• Quickly look up how to operate OpenClaw Ollama Runtime.
• Reduced command errors between native Ollama (/api/*) and OpenAI-compatible (/v1/*) endpoints.
• Provides a consistent debug checklist in a modular Android environment.

When to use

• Remember the command start/stop/restart/status/log.
• Need to pull new model and verify model works.
• Error 401/404, stale PID, or healthcheck failed.
• Need to confirm the standard path of the module/runtime before operating.

Process of using skills

1. Read references/usage-cheatsheet.md to select the correct command.
2. If there is an error: read references/troubleshooting.md according to symptom.
3. When manipulating the model: read references/model-management.md.
4. If runtime status is unknown: run scripts/diag.sh.

Guardrails

• This skill is for reference only; Do not patch the code/module yourself.
• Do not operate outside the Ollama runtime unless explicitly requested.
• Do not write secrets/tokens into KB.

Output contract

When using this skill to reply to users, summarize in the following format:

openclaw_ollama_kb:
  intent: <start|stop|status|log|pull|debug>
  commands:
    - <command 1>
    - <command 2>
  notes:
    - <important caveat>
  followup:
    - <optional next step>

Trigger Rules

• Positive: Triggers when the user request matches openclaw-ollama-runtime-kb according to description: Knowledge Base looks up how to use OpenClaw Ollama Runtime on Android (start/stop/status/log, pull model, quick debug).
• Negative: Not enabled for requests outside scope or without the workflow/expertise of this skill.
• Escalation: If there is an external action/network write or sensitive operation, an approval checkpoint must be added before executing.

Output Contract

• Deliverable: target tracking result of openclaw-ollama-runtime-kb; Specify the artifact/path/steps performed (if any).
• Validation: reports PASS/FAIL status and main checks have been run.
• Failure mode: when there is an error, the root cause + proposed fix must be stated.

Error Handling

• If input/permissions are missing: report what is missing and clearly request additional information.
• If runtime/tool ​​error: brief error return + possible cause + suggested fix.
• If there is a risk policy/safety: stop and escalate according to approval checkpoint.