Use ONLY when the user explicitly asks Claude to delegate, hand over, pass, or send work to Codex CLI - phrases like "delega a codex", "usa codex", "handover a codex", "passa a codex", "ask codex", "@codex", "/codex", "via codex", or "with codex". Runs `codex exec` non-interactively and returns Codex's final answer verbatim. NEVER invoke this skill unless the user explicitly names Codex or Codex CLI as the delegate.
Use ONLY when the user explicitly asks to delegate a task to GitHub Copilot CLI — phrases like "delega a copilot", "usa copilot", "ask copilot", "passa a copilot", "fai con copilot", "@copilot", "/copilot", "via copilot", "with copilot". Runs `copilot -p` non-interactively, defaulting to Claude Opus 4.7, so the token cost of large/expensive work (big-file reads, broad explanations, shell-output analysis) bills against the user's GitHub Copilot subscription instead of their Claude subscription. Acts as a transparent passthrough: build a full prompt with the needed context, execute `copilot`, return the response verbatim inside a fenced block. NEVER invoke this skill on your own — only when the user explicitly names Copilot or uses the trigger phrases above. If the user just asks to read a big file or run a long command without naming Copilot, do NOT trigger this skill.
Use ONLY when the user explicitly asks Claude to delegate, hand over, pass, or send work to Gemini CLI — phrases like "delega a gemini", "usa gemini", "passa a gemini", "fai con gemini", "ask gemini", "@gemini", "/gemini", "via gemini", "with gemini". Runs `gemini -p` non-interactively and returns Gemini's answer verbatim. The point is to offload token-heavy work (long file analysis, broad codebase Q&A, big log summaries) to the user's Gemini account quota instead of the Claude subscription. NEVER invoke this skill unless the user explicitly names Gemini or Gemini CLI as the delegate.
Use ONLY when the user explicitly asks the AI assistant to route, dispatch, fan-out, split, or parallelize a task across multiple AI model CLIs (Claude / Codex / Gemini / Copilot's many models) — phrases like "usa lo swarm", "swarm:", "/swarm", "instrada", "fan-out", "route this task", "scegli il modello migliore per", "delega allo swarm", "best model for this", "ensemble", "second opinion da", "team mode", "parallel-dev", "sviluppa in parallelo", "swarm parallel", "fai team", "build with a team". Four modes: single dispatch, sequential pipeline, parallel fan-out, and parallel-team. Works across any harness that can read this SKILL.md (Claude Code, Copilot, Gemini CLI, Codex) — the dispatch mechanism differs per harness (see Harness-specific dispatch). Hierarchical orchestration only — there is NO peer-to-peer messaging between different model CLIs. NEVER invoke this skill on your own; only when the user explicitly uses one of the trigger phrases above.
any input (code, docs, papers, images, videos) to knowledge graph. Use when user asks any question about a codebase, documents, or project content - especially if graphify-out/ exists, treat the question as a /graphify query.
any input (code, docs, papers, images, videos) to knowledge graph. Use when user asks any question about a codebase, documents, or project content - especially if graphify-out/ exists, treat the question as a /graphify query.
Produces and incrementally maintains a versioned reverse-engineering dossier for a subject — a legacy codebase, a microservice, a third-party API, an internal library, a data pipeline. Output is markdown chapters under `docs/support/reverse/<subject-slug>/`, combining project documents (sprint inputs, meeting notes, functional specs) with the subject's code/artifacts indexed via the graphify knowledge graph. Flow — (1) graphify indexing loads context only, (2) long-running interactive Q&A on graph and source, (3) explicit save trigger ("save", "ok write it", "make v1.5") bumps minor version, updates only chapters in the consolidated intent, appends a changelog entry. Triggers — "reverse-engineer X", "document legacy Y", "build a dossier on vendor Z's API", "refresh the reverse on <subject>", "new version of the reverse", or mention of a path under `docs/support/reverse/` in an update context. Do NOT use for one-shot graph queries or single-chapter manual edits.
Acceptance Test Runner — reads acceptance criteria from a sprint manifest and/or requirements files, generates a machine-readable testbook, executes it via Playwright against the running app, produces markdown reports with screenshots, and maintains a backlog of failures. Optionally pushes failures to an external tracker (Paseo integration planned as future work). Use this skill when the user mentions acceptance testing, testbook, running tests against the app, producing a stakeholder-friendly test plan, verifying the app against the sprint manifest, or pushing test failures to a tracker. Triggers — `/atr generate`, `/atr run`, `/atr doc`, `/atr push`, "run acceptance tests", "generate testbook for sprint N", "check the app against the manifest", "create a readable test plan for stakeholders". Language — reports and testbook descriptions follow the project's working language (inferred from the sprint manifest / requirements content, or from `nonoise.config.json` `language` if set; defaults to English).