intent-analyzer

---

LIBRARY-FIRST PROTOCOL (MANDATORY)

Before writing ANY code, you MUST check:

Step 1: Library Catalog

• Location: .claude/library/catalog.json
• If match >70%: REUSE or ADAPT

Step 2: Patterns Guide

• Location: .claude/docs/inventories/LIBRARY-PATTERNS-GUIDE.md
• If pattern exists: FOLLOW documented approach

Step 3: Existing Projects

• Location: D:\Projects\*
• If found: EXTRACT and adapt

Decision Matrix

Match	Action
Library >90%	REUSE directly
Library 70-90%	ADAPT minimally
Pattern exists	FOLLOW pattern
In project	EXTRACT
No match	BUILD (add to library after)

---

STANDARD OPERATING PROCEDURE

Purpose

• Clarify what the user actually needs before executing research work.
• Separate HARD / SOFT / INFERRED constraints and validate them.
• Prepare downstream skills with structured context and explicit ceilings.

Trigger Conditions

• Positive: unclear or under-specified research requests; conflicting goals; prioritization questions.
• Negative: clear prompt-optimization tasks (route to prompt-architect) or full orchestration (use deep-research-orchestrator).

Guardrails

• Two-pass refinement: (1) capture/structure signals; (2) epistemic check with confidence ceilings.
• Avoid inventing goals; label INFERRED items and request confirmation if possible.
• Maintain structure-first outputs: SKILL + README + examples + references updated when patterns emerge.

Inputs

• Raw request or prompt snippet.
• Known context (domain, audience, deadlines, format).

Workflow

1. Signal Extraction: Pull verbs, objects, audience, deliverable types, timelines, and exclusions.
2. Constraint Bucketing: Sort into HARD / SOFT / INFERRED with evidence/source labels.
3. Intent Hypothesis: Draft the likely intent, success criteria, and anti-goals.
4. Validation Pass: Check for ambiguity, conflicts, or missing constraints; propose clarifying questions.
5. Handoff Prep: Produce a concise brief for downstream skills with confidence ceiling stated.

Validation & Quality Gates

• INFERRED items explicitly marked and minimized.
• Confidence ceiling included; avoid overclaiming beyond inference/report.
• Output is English-only and structured for reuse.

Response Template

**Intent Hypothesis**
- Primary: ...
- Secondary: ...

**Constraints**
- HARD: ...
- SOFT: ...
- INFERRED (confirm): ...

**Success / Anti-Goals**
- ...

**Clarifying Questions**
- ...

Confidence: 0.76 (ceiling: inference 0.70) - based on parsed request signals.

Confidence: 0.76 (ceiling: inference 0.70) - Default after completing the SOP.