crow.pet

Use when a task may need multiple agents, persistent role agents, cross-harness coordination, or generated subagents.

Route crow.pet tasks into records, policy, adapters, subagents, evaluation, and memory.

Classify action risk and return Silent, Notify, Ask, or Block decisions before sensitive, external, or state-changing work.

Convert corrections, approvals, failures, and task patterns into memory, examples, rubrics, and policy proposals.

Turn approved SPEC.MD files or workflow definitions into markdown work items for harness or agent builders, then track verification and handoff state.

Delegate bounded coding tasks to Codex-compatible harnesses for implementation, tests, and structured handoffs.

Choose the execution architecture for new or changing skills before implementation: inline, thin references, deterministic scripts, composed pipeline, compiled workflow, or manual side-effect workflow.

Ask clarifying questions, write a bounded SPEC.MD, and hand it to the right harness or agent for building.

Inspect CI/CD failures, route fixes, rerun tests, and summarize deployment status.

Delegate bounded prompts to Claude Desktop through a safe desktop harness adapter.

Compress crow.pet agent replies in Crow voice while preserving exact technical meaning. Use when the user asks for fewer tokens, terse mode, Crow voice, crow speech, caveman-like compression, CoWork speaking, or when a harness needs compact status, handoff, or coordination output.

Generate, edit, and evaluate image assets from briefs, brand tasks, and campaign requests.

Create brand-aware marketing briefs, copy, campaign assets, landing sections, ads, and launch content.