Legal Briefing
npx machina-cli add skill fedec65/bettercallclaude/legal-briefing --openclawLegal Briefing Skill
You are aware of BetterCallClaude's briefing session capability. When you detect a complex legal query that would benefit from structured intake before execution, suggest or activate a briefing session.
Activation Criteria
Suggest a briefing session when any of the following conditions are met:
Complexity Indicators (score >= 5)
- Three or more legal domains in a single query (e.g., corporate + tax + compliance).
- Multi-jurisdictional question (federal + cantonal, or multiple cantons).
- Document output expected alongside analysis (e.g., "analyze and draft").
- Financial exposure mentioned above CHF 100,000.
- Multiple parties with conflicting interests.
Ambiguity Indicators
- Desired output is unclear (research? strategy? document? compliance check?).
- Jurisdiction cannot be determined from the query.
- Multiple valid routing paths exist with significantly different outcomes.
- The query uses vague terms: "handle", "deal with", "figure out", "advise on".
Pipeline Indicators
- Query maps to a workflow template (litigation-prep, due-diligence, contract-lifecycle).
- Three or more agents would need to be coordinated.
- Data dependencies exist between agents (output of one feeds into another).
When NOT to Activate
Skip briefing and route directly when:
- Query is a simple, focused question (complexity 1-3).
- User explicitly specifies a single agent with
@agent_name. - User uses
--skip-briefingor--directflag. - Query is a citation lookup, translation, or format verification.
- User is resuming a previously approved execution plan.
Activation Behavior
Implicit Suggestion (auto-detected complexity >= 5)
When the /legal gateway scores a query at complexity 5+ and no --skip-briefing flag is present:
š” This query involves [N domains / multi-jurisdiction / pipeline coordination].
A briefing session would help build a precise execution plan before I start working.
**Options**:
- **Proceed with briefing** (recommended) ā I'll ask a few targeted questions first
- **Skip briefing** ā Route directly to agents (add `--skip-briefing` to bypass in future)
Explicit Activation
When the user invokes /bettercallclaude:briefing, always activate regardless of complexity.
Memory Key Patterns
The briefing system uses these memory keys for persistence:
| Pattern | Purpose |
|---|---|
briefing_[id] | Full state for a specific briefing session |
briefing_latest | ID of the most recently active briefing |
briefing_index | Registry of all briefing sessions with status |
Integration Points
/legalgateway: Briefing phase is embedded between intent classification and routing.- Orchestrator agent: Consumes the approved execution plan YAML for step-by-step execution.
- All specialist agents: Can be spawned as panel members during the consultation phase.
- Summarizer agent: Can be appended to the execution plan's final stage.
Quality Standards
- Briefing suggestions must be helpful, not obstructive ā never block simple queries.
- Auto-detection should have a low false-positive rate: only suggest for genuinely complex cases.
- The user must always have a clear path to skip the briefing if they prefer direct routing.
Source
git clone https://github.com/fedec65/bettercallclaude/blob/main/bettercallclaude/skills/legal-briefing/SKILL.mdView on GitHub Overview
Legal Briefing auto-activates structured intake for complex legal queries to build a precise execution plan before work begins. It detects criteria like multi-domain, multi-jurisdiction, and outputs such as analyzes and drafts, then suggests or activates a briefing session. This helps coordinate agents and ensure a clear path to resolution.
How This Skill Works
The skill evaluates a query against activation criteria and assigns a complexity score. If the score is 5 or higher and no --skip-briefing flag is present, it suggests a briefing session; users can accept or decline. Explicit activation via /bettercallclaude:briefing always starts the briefing. It uses memory keys like briefing_[id], briefing_latest, and briefing_index to persist state and integrates with the Orchestrator to produce an execution plan YAML for multi-agent coordination.
When to Use It
- Query includes three or more legal domains in a single request (e.g., corporate, tax, and compliance).
- The question is multi-jurisdictional (federal plus cantonal or multiple cantons).
- The output required includes both analysis and a drafted document (e.g., analyze and draft).
- Three or more agents would need coordination or there are data dependencies between agents.
- Ambiguity about the desired output or multiple valid routing paths exist.
Quick Start
- Step 1: When the /legal gateway scores a query at complexity 5+ and no --skip-briefing flag is present, consider a briefing.
- Step 2: If the user agrees, initiate a briefing session and ask targeted intake questions to scope the plan.
- Step 3: Generate and review the execution plan YAML, then route to the appropriate agents for coordinated execution.
Best Practices
- Use briefing only for genuinely complex cases (complexity score of 5 or higher).
- Ask targeted intake questions before proceeding to build a precise plan.
- Clearly define the expected outputs to tailor the briefing (e.g., deliverables or milestones).
- Review the generated execution plan YAML before routing to agents.
- Provide an option to skip briefing if the user prefers direct routing (--skip-briefing or --direct).
Example Use Cases
- A multinational corporation requests cross-border tax, corporate governance, and regulatory compliance analysis.
- A complex litigation matter requiring a litigation-prep workflow across federal and state laws with a draft complaint.
- A due-diligence review mapping data dependencies across legal, financial, and contract teams.
- A cross-border M&A scenario needing multi-jurisdictional analysis and contract lifecycle planning with drafts.
- A contract renegotiation involving conflicting interests among multiple parties and coordinated inputs from several specialists.