researchlisted

# Research Skill Orchestrator-driven deep research. The skill plans the work itself, spawns parallel deep-research subagents per topic cluster, synthesizes findings under unbounded review, and produces a polished report. --- ## Auth Approach No authentication required. Uses the host agent's web search and fetch or browse capability, and writes to the local filesystem. ## Tool Preference 1. **Subagent dispatch when available and permitted**: dispatch researcher, synthesis, writer, and reviewer agents. 2. **File read tools**: load prompt templates before dispatch. 3. **Shell**: directory creation, date generation, simple file globbing for verification. 4. **Progress list**: live loop bookkeeping. 5. **Web search and fetch or browse tools**: fallback only if subagent dispatch fails. ## Platform Adaptation Use the host platform's equivalent tools without changing the workflow: | Capability | Claude Code | Codex | |---|---|---| | Subagent dispatch | Agent tool | `spawn_agent` only when available and permitted. Otherwise run the phase inline. | | Progress list | TaskCreate, TaskUpdate, TaskList | `update_plan` | | Web research | WebSearch, WebFetch | `web.run` search and open calls, or the host browser/search tools | | File reads | Read | shell reads such as `sed`, `rg`, or equivalent file read tools | | Shell | Bash | shell command tool | When a platform cannot dispatch subagents for the current request, keep the same artifact boundaries and run each phase inline in the