Menu
Delegate a self-contained task to a verified cheaper/faster child-Pi model in a fresh ephemeral context.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Use the stealth browser for human-like web automation against bot-protected sites. Handles login flows, scraping, and form filling with anti-detection bypass.
Export a chat timeline to a PDF using the internal localhost export endpoint and wkhtmltopdf.
Nightly Bayesian interaction-quality classifier — flags behavioral patterns from chat history and writes self-improvement reflections without spending model tokens on classification.
Record major friction or postmortem feedback in the configured workspace vent log.
Open a spreadsheet-style widget for a Markdown table so the user can edit it and send the final Markdown table back into chat.
git_history and json_query tools for structured git log exploration and jq-style JSON querying.
| name | delegate |
| description | Delegate a self-contained task to a verified cheaper/faster child-Pi model in a fresh ephemeral context. |
| distribution | public |
Use the delegate tool for self-contained work that does not need the conversation history. The child runs with a verified executable model, a bounded tool profile, and no persistent Pi session.
judgeDo not delegate when the task depends on conversation history, needs user interaction, requires frontier capability, or is simpler than describing a complete standalone prompt.
delegate({
prompt: "Read src/client.ts. Summarize its public API in five bullets and identify compatibility risks.",
files: ["src/client.ts"],
task_category: "summarize",
tools: "read_only"
})
| Category | Target | Use for |
|---|---|---|
quick | Tier 2 | Formatting, extraction, translation, factual Q&A |
summarize | Tier 2 | File, note, and code summaries |
code | Tier 3 | Code generation and refactoring |
analyze | Tier 3 | Review, architecture, and debugging |
reason | Tier 3 | Complex planning and multi-step logic |
judge | Tier 3 | Critical review from another family when available |
Automatic selection is capped at the verified tier of the current model. If the current model is unknown to Delegate policy, automatic selection fails closed.
Use model: "provider/model" only when an exact executable child-CLI model is required. Overrides bypass automatic tier and configured provider/model exclusions, but they do not bypass executability or image-capability checks. A model visible only to Piclaw runtime is not enough.
| Profile | Child tools |
|---|---|
read_only | read,grep,find,ls,mcp |
standard (default) | read,grep,find,ls,bash,mcp |
full | read,grep,find,ls,bash,edit,write,mcp |
A custom comma-separated list of Pi child built-ins is accepted. Do not assume Piclaw add-on tools are present: the child loads only the discovered MCP adapter in addition to core tools.
/workspace; symlink escapes and non-regular files are rejected.When the user asks to double-check, verify, review, or obtain a second opinion:
delegate({
prompt: "Review the response below for correctness, omissions, and unsupported claims. Be critical.\n\n<response>\n...\n</response>",
task_category: "judge",
tools: "read_only"
})
Judge selection crosses families only when a valid tier-safe executable alternative exists.
The child has no conversation history. Include all necessary facts, paths, constraints, and the desired output shape.
"Fix the bug we discussed.""Read src/cache.ts and tests/cache.test.ts. Diagnose why stale entries survive invalidate(), propose the smallest fix, and return a patch outline."Use the final result normally. While it runs, Delegate may emit bounded structured status/tool progress. It retries automatically only for provider setup, authentication, or unavailable-model failures; ordinary execution, protocol, timeout, cancellation, and rate-limit failures are returned without retry.