| name | animal-handling |
| description | Use this skill when the system is not fully broken but behaves like a skittish animal: flaky CI, fragile legacy services, or moody dependencies. It focuses on coaxing, pacing, and stable handling rather than heroic rewrites. |
| user-invocable | true |
Animal Handling
Calm a temperamental system long enough to get useful work out of it.
Overview
Animal Handling is interpreted here as a metaphorical skill with a shipping-now execution model.
Canonical source: Animal Handling (skill)
Provider target: OpenClaw
When To Use
- A flaky or non-deterministic system keeps reacting badly to normal operations.
- Maintenance work depends on reading triggers, pacing changes, and not spooking the stack.
- You need a handling plan that reduces agitation before deeper repair work.
Workflow
- Observe the failure patterns, triggers, and safe handling rules the system seems to respond to.
- Choose gentle interventions, sequencing, and retries that reduce instability instead of amplifying it.
- Return a stability playbook plus the signs that the system is calming down or getting worse.
Deliverables
- A system-handling plan tuned to the temperamental behavior.
- A list of triggers, soothing interventions, and escalation signals.
- Criteria for when to continue, pause, or hand off to a deeper fix.
Guardrails
- Do not replace diagnosis with superstition; name which patterns are evidence and which are hunches.
- Avoid making the system more dependent on manual soothing without documenting root-cause work.
Default Invocation
Use $animal-handling to steady this temperamental system and tell me how to work with it without provoking a failure.
