From brief
Product context, decisions, and strategic reasoning via the Brief MCP server. Use when starting a task, making architectural decisions, or needing product knowledge.
How this skill is triggered — by the user, by Claude, or both
Slash command
/brief:briefThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Brief holds the business context, customer insights, and strategic decisions that live outside any single document or conversation. It exposes that knowledge to this agent through the Brief MCP server.
Brief holds the business context, customer insights, and strategic decisions that live outside any single document or conversation. It exposes that knowledge to this agent through the Brief MCP server.
This skill teaches you when to call Brief MCP tools. The tool descriptions surfaced during the MCP handshake teach you how to call each one.
Call brief_get_onboarding_context once at the beginning of a session. It returns a structured snapshot: product strategy, personas, active decisions, recent work, and velocity.
scope: 'full' returns the complete strategic snapshot.scope: 'product' returns just the product context object — narrower and faster.scope: 'identity' returns just workspace id and name — cheap sanity check.This is the canonical first-contact tool. Do not use brief_search or brief_ask as a substitute — search is for finding one specific thing, ask is for strategic reasoning after context is loaded.
If Brief has no context yet, send the user's framing to brief_ask with onboarding: true and a bounded context payload: prioritize the package manifest + git log --oneline -20, then add a trimmed README excerpt if space allows. Brief synthesizes product context and responds with its first strategic insight.
Call brief_ask with mode: "check" and the proposed direction phrased as a question. Brief validates the approach against existing decisions, customer signals, and team priorities, and flags conflicts.
Example: brief_ask({ question: "Should we use Postgres or DynamoDB for primary storage?", mode: "check" }).
Call brief_ask with mode: "advise" for guidance grounded in your org's real data — past decisions, customer feedback, current priorities.
For sub-2-second context checks mid-build, use mode: "fast".
When a consequential choice is made, record it via brief_record_decision with the decision and rationale. Brief renders a decision approval card so the user can approve, edit, or reject before it is persisted — this is the preferred path.
For rare batch or escape-hatch writes, fall back to brief_execute_operation with operation: "record_decision".
Do not store product decisions in conversation memory. Conversation memory is local and ephemeral; Brief is shared and permanent.
For factual retrieval — finding specific decisions, documents, personas, features — call brief_search with the appropriate types filter, or brief_browse for structured listings. Use brief_list_pending_decisions for decisions awaiting approval.
Use brief_ask for judgment calls, not for data retrieval.
Capture conversation_id from the first brief_ask response and pass it to subsequent calls in the same thread so Brief maintains context across turns.
Call brief_update_product_context to write company, goals, or persona fields as the user shares them. Prefer silent updates during onboarding over announcing each write.
| Tool | Purpose |
|---|---|
brief_get_onboarding_context | Primer — load full workspace snapshot (decisions, personas, features, pipeline). Call once at session start. |
brief_ask | Strategic reasoning, multi-turn, grounded in your org's data. Modes: advise, check, fast. |
brief_search | Keyword search across decisions, documents, personas, features, signals. Filter with types. |
brief_browse | Structured listings of workspace objects. |
brief_record_decision | Capture a decision with rationale — renders an approval card. |
brief_confirm_decision / brief_reject_decision / brief_archive_decision | Lifecycle actions on decisions. |
brief_list_pending_decisions | Decisions awaiting approval. |
brief_update_product_context | Write company, goals, or persona fields. |
brief_execute_operation | Escape hatch for operations without a direct tool — prefer direct tools when available. |
brief_report_progress / brief_report_completion | Send phase updates and completion status for in-flight coding tasks. |
If the MCP host supports prompts, the server also registers brief-setup, brief-welcome-back, and brief-context. Users can invoke these from the host's prompt picker as shortcuts for the flows above.
brief_get_onboarding_context once at session start before reasoning about the user's product.brief_ask for judgment; use brief_search / brief_browse for factual lookups.conversation_id from the first brief_ask response and reuse it on subsequent calls in the same thread.workspace_url and per-object view_url fields from tool responses.brief ask / brief context / brief decisions — that is the CLI channel, which lives behind a different install path (brief init).npx claudepluginhub brief-hq/brief-plugins --plugin briefProvides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Fetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.