seltz
Use Seltz's official MCP search tool for current, source-backed web research in Codex.
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-14
- Stars
- ★ 1.6k
- Forks
- 491
01 前言 Front Matter
- name
- seltz
- license
- Apache-2.0
- description
- Use Seltz's official MCP search tool for current, source-backed web research in Codex.
03 SKILL.md 正文
Seltz web research
Use the configured Seltz MCP server when the user needs current web information, recent developments, source discovery, or a fact check. The server exposes Seltz's documented tools for web search and page fetching; take their parameters and limits from the live tool definitions, which are documented on the Seltz MCP page.
Operating rules
- Turn the user's request into a concise, specific query. Include a date, jurisdiction, product, or domain when it materially narrows the research question.
- Request only as many results as the task needs. Ask for more only when the first results leave a clear evidence gap.
- Preserve source URLs and publication dates from the tool result. Distinguish Seltz output from your own synthesis, and tell the user when a claim is time-sensitive or only supported by one source.
- Prefer primary, official, or directly attributable sources for legal, financial, medical, security, and product claims. Do not treat ranking as proof.
- Do not send API keys, passwords, payment credentials, private account data, or unnecessary personal information in a query.
- Do not imply that Seltz verified a fact or returned an exhaustive result set unless the live tool response explicitly supports that statement. When the contents of a specific page matter, fetch that page instead of quoting a search snippet.
Failure handling
- Authentication or permission failures: ask the user to reconnect Seltz in Codex and sign in again; never ask them to paste a key or other credential into the conversation.
- Rate, usage, or balance limits: report the limit, honor any retry timing returned by the server, and do not loop.
- Timeouts or unavailable service: retry at most once when the request is safe and idempotent, then offer a fallback that does not pretend to have current Seltz data.
- Malformed or incomplete results: say what is missing, preserve any usable source links, and do not manufacture fields or citations.
Safe response pattern
For research answers, state the question and retrieval time, summarize the strongest sources, link them, and call out uncertainty, conflicting evidence, or missing coverage. For fact checks, separate the verdict from the evidence and explain what would change the conclusion.
04 同仓库更多 Skills
worktable-review-with-annotations
Create, inspect, reply to, update, or resolve situated Worktable annotations. Use when feedback belongs to a Doc section, block, text range, or HTML Doc, or when the user asks to address existing review feedback and coordinate a requested artifact edit.
worktable-find-and-synthesize-context
Find, assess, and synthesize context from Worktable Docs and Records. Use when the user asks what a workspace contains, wants workspace-backed research, needs current facts reconciled across artifacts, or requests a cited synthesis without changing workspace state.
worktable-create-or-update-docs
Create, edit, rename, or maintain narrative Worktable Docs. Use for plans, decisions, briefs, notes, research syntheses, and other durable prose when the user requests a new artifact or a safe update to an existing one.
worktable-create-or-manage-records
Organize repeated structured information as durable Worktable Records. Use for research datasets, inventories, directories, pipelines, trackers, and other collections whose items need identity, fields, filtering, links, validation, or individual updates.
worktable-create-or-manage-interactive-html
Create or improve durable visual and interactive artifacts in Worktable. Use for dashboards, trackers, calculators, mini apps, slide presentations, and other requests that benefit from a polished HTML Doc.
worktable-collaborate-threads
Read, answer, and hand off messages through durable Worktable threads with users or other authenticated participants. Use when the user asks to check addressed messages, reply to a participant, continue a thread, request another agent's help, wait briefly for a response, or watch for comments and messages sent to the agent.