delegate-to-ollama
by IsmaelMartinez
Use this skill to offload non-reasoning text work to locally-installed Ollama models via ollama run, keeping content on-device and freeing the main-agent context window. Saves API tokens too, though context protection and privacy are the headline values, not cost. MUST use whenever the user asks to summarise a log/diff/file/PR/issue, draft a commit message/changelog/release note, triage or classify many items, extract structured fields from free text, skim many files for a one-liner, rewrite or reformat prose, anonymise or redact text, convert between markup formats (YAML↔JSON, markdown→HTML), generate regex from description, or stub docstrings. MUST also use when the user mentions running locally, on-device, offline, saving tokens, privacy, Ollama, or a local model. MUST also use after the user sets durable auto-delegate intent ("delegate where it fits", "auto-delegate", "route to ollama as appropriate"), defaulting to delegating any subsequent matching task without re-confirming. Do NOT use for code correctness review, architectural decisions, debugging or tracing errors, implementing features, or any task whose output triggers a destructive or shared-state action without review. Do NOT use for open-ended "find anything interesting / suggest improvements" prompts that ask the model to surface things not in the input — those invite fabrication.