dew-method 0.1.0

package/src/dew-skills/1-analysis/dew-decision-gate/assets/gate-catalog.md

@@ -0,0 +1,412 @@
+# DEW Decision Gate Catalog
+
+## Phase 00 — Project Framing
+
+### GATE-00 — Project Scope Gate
+
+Decision:
+- learning project
+- portfolio project
+- internal decision-support
+- production-grade
+
+Required before:
+- Data Product PRD
+
+Related HALT:
+- HALT-00
+- HALT-01
+
+---
+
+## Phase 01 — Business, KPI & Data Product Discovery
+
+### GATE-01 — Data Consumer Gate
+
+Decision:
+- who consumes the data product
+
+Required before:
+- KPI discovery
+- source discovery
+
+Related HALT:
+- HALT-01
+- HALT-15
+
+### GATE-02 — Business Decision Gate
+
+Decision:
+- what decision the data product supports
+
+Required before:
+- KPI definition
+
+Related HALT:
+- HALT-01
+
+### GATE-03 — KPI Definition Gate
+
+Decision:
+- accepted KPI definition
+- formula
+- grain
+- dimensions
+- required fields
+
+Required before:
+- KPI feasibility validation
+
+Related HALT:
+- HALT-02
+
+### GATE-04 — KPI Grain Gate
+
+Decision:
+- one row represents what
+
+Required before:
+- KPI feasibility
+- data model design
+
+Related HALT:
+- HALT-07
+
+### GATE-05 — Trust Expectation Gate
+
+Decision:
+A. Exploratory only
+B. Internal decision-support with documented limitations
+C. Trusted shared product with quality checks
+D. Production-grade with contracts and monitoring
+
+Required before:
+- source priority
+- architecture
+- serving design
+
+Related HALT:
+- HALT-08
+
+---
+
+## Phase 02 — KPI Feasibility Validation
+
+### GATE-06 — KPI Feasibility Gate
+
+Decision:
+- verified
+- feasible with caveats
+- hypothesis only
+- rejected
+- deferred
+
+Required before:
+- architecture
+
+Related HALT:
+- HALT-03
+- HALT-18
+
+---
+
+## Phase 03 — Source Discovery & Validation
+
+### GATE-07 — Source Access Gate
+
+Decision:
+- access verified
+- access blocked
+- manual only
+- reject source
+
+Required before:
+- source validation
+- source priority
+
+Related HALT:
+- HALT-05
+
+### GATE-08 — Source Trust Gate
+
+Decision:
+A. P1 — Use in MVP
+B. P2 — Use later
+C. Exploration only
+D. Reject
+
+Required before:
+- architecture
+- ingestion design
+
+Related HALT:
+- HALT-06
+
+### GATE-09 — Source Grain Gate
+
+Decision:
+- source grain approved
+- needs bridge/mapping
+- incompatible with target KPI grain
+- reject/defer source
+
+Required before:
+- transformation design
+
+Related HALT:
+- HALT-07
+
+### GATE-10 — Source Freshness Gate
+
+Decision:
+- freshness sufficient
+- freshness insufficient but acceptable with caveat
+- source needs alternate update process
+- reject source
+
+Required before:
+- ingestion frequency design
+
+Related HALT:
+- HALT-09
+
+### GATE-11 — Schema Evolution Gate
+
+Decision:
+A. Strict schema contract
+B. Schema snapshot + drift detection
+C. Flexible rescued fields
+D. Manual review on change
+
+Required before:
+- ingestion design
+
+Related HALT:
+- HALT-13
+
+---
+
+## Phase 04 — Requirement Gate
+
+### GATE-12 — Architecture Readiness Gate
+
+Decision:
+A. Proceed to architecture
+B. Proceed with caveats
+C. Return to KPI validation
+D. Return to source validation
+E. Reduce MVP scope
+
+Required before:
+- data architecture
+
+Related HALT:
+- HALT-10
+
+---
+
+## Phase 05 — Data Architecture
+
+### GATE-13 — Architecture Pattern Gate
+
+Decision:
+- local-first
+- warehouse-first
+- lakehouse
+- hybrid
+- streaming-first
+
+Required before:
+- storage design
+
+Related HALT:
+- HALT-10
+
+### GATE-14 — Technology Choice Gate
+
+Decision:
+- tool/platform choice
+- rejected alternatives
+- revisit trigger
+
+Required before:
+- implementation stories
+
+Related HALT:
+- HALT-17
+
+---
+
+## Phase 06 — Storage & Ingestion Design
+
+### GATE-15 — Storage Layer Gate
+
+Decision:
+- raw/bronze/silver/gold shape
+- metadata expectations
+- retention
+
+Required before:
+- ingestion implementation
+
+Related HALT:
+- HALT-11
+
+### GATE-16 — Ingestion Pattern Gate
+
+Decision:
+- pull API
+- file drop
+- database export
+- CDC
+- stream
+- manual
+
+Required before:
+- ingestion story
+
+Related HALT:
+- HALT-12
+
+### GATE-17 — Batch vs Streaming Gate
+
+Decision:
+- batch
+- micro-batch
+- streaming
+- event-driven
+
+Required before:
+- orchestration design
+
+Related HALT:
+- HALT-12
+
+---
+
+## Phase 07 — Transformation, Modeling & DQ Design
+
+### GATE-18 — Data Quality Rule Gate
+
+Decision:
+- minimum DQ rules
+- warning vs blocking
+- quarantine behavior
+
+Required before:
+- gold modeling
+- serving
+
+Related HALT:
+- HALT-14
+
+### GATE-19 — Transformation Rule Gate
+
+Decision:
+- business rules approved
+- rule caveats
+- deferred logic
+
+Required before:
+- data story implementation
+
+Related HALT:
+- HALT-14
+
+---
+
+## Phase 08 — Serving, Governance & DataOps
+
+### GATE-20 — Serving Mode Gate
+
+Decision:
+- dashboard
+- data mart
+- API
+- ML feature table
+- notebook/report
+- reverse ETL
+
+Required before:
+- serving implementation
+
+Related HALT:
+- HALT-15
+
+### GATE-21 — Privacy Posture Gate
+
+Decision:
+- no sensitive data
+- low-risk internal
+- PII present with controls
+- high-risk requires governance review
+
+Required before:
+- implementation
+
+Related HALT:
+- HALT-16
+
+### GATE-22 — Operational Readiness Gate
+
+Decision:
+- minimal logs
+- monitored internal pipeline
+- SLA-backed production pipeline
+- not operationally ready
+
+Required before:
+- production-like serving
+
+Related HALT:
+- HALT-17
+
+---
+
+## Phase 09 — Epics, Stories & Readiness
+
+### GATE-23 — Implementation Readiness Gate
+
+Decision:
+A. Ready for stories
+B. Ready with caveats
+C. Return to architecture
+D. Return to source/KPI validation
+
+Required before:
+- sprint planning
+
+Related HALT:
+- HALT-17
+
+---
+
+## Phase 10 — Build, Review & Retrospective
+
+### GATE-24 — Validation Evidence Gate
+
+Decision:
+- evidence accepted
+- evidence incomplete
+- rerun validation
+- revise story
+
+Required before:
+- story done
+
+Related HALT:
+- HALT-18
+
+### GATE-25 — Learning Capture Gate
+
+Decision:
+- learning captured
+- pattern extracted
+- needs retrospective
+- close epic
+
+Required before:
+- epic closure
+
+Related HALT:
+- HALT-19

package/src/dew-skills/1-analysis/dew-decision-gate/assets/gate-frontmatter-schema.md

@@ -0,0 +1,20 @@
+# Decision Gate Frontmatter Schema
+
+Artifacts that participate in DEW decision gates should include frontmatter.
+
+## Minimal schema
+
+```yaml
+---
+title: "{Artifact Title}"
+status: draft
+created: YYYY-MM-DD
+updated: YYYY-MM-DD
+phase: "{phase-id}"
+workflow: dew
+stepsCompleted: []
+decisionGates: {}
+evidence: []
+assumptions: []
+caveats: []
+---

package/src/dew-skills/1-analysis/dew-decision-gate/assets/global-rules.md

@@ -0,0 +1,111 @@
+# DEW Global Rules
+
+## Core Rule 1 — Agent may recommend, but must not decide
+
+The agent may analyze, explain, compare options, and recommend one option.
+
+The agent must not silently make project-shaping decisions for the user.
+
+Project-shaping decisions include:
+- project scope
+- business objective
+- data consumer
+- KPI definition
+- KPI grain
+- KPI feasibility
+- source acceptance
+- source priority
+- trust expectation
+- freshness
+- architecture pattern
+- technology choice
+- storage layer
+- ingestion pattern
+- schema evolution
+- transformation rule
+- data quality rule
+- serving mode
+- privacy/security posture
+- implementation readiness
+- learning closure
+
+For every project-shaping decision, the agent must:
+1. Stop.
+2. Explain the concept briefly.
+3. Show current evidence.
+4. Present options.
+5. Recommend one option.
+6. Explain trade-offs.
+7. Ask the user to choose.
+8. Wait for explicit user confirmation.
+9. Record the decision in `.decision-log.md`.
+
+The agent may recommend, but must not decide.
+
+## Core Rule 2 — No phase is done without evidence
+
+A phase is not complete just because an artifact was written.
+
+A DEW phase is complete only when it has:
+- artifact
+- validation evidence
+- decision log entry
+- learning note
+- next-phase readiness statement
+
+Planning artifacts are not evidence.
+
+Evidence may include:
+- successful API call log
+- downloaded source sample
+- schema snapshot
+- profiling report
+- executable SQL/Python notebook
+- source contract test
+- data quality test result
+- sample KPI output
+- validation report
+- user-approved decision log entry
+
+If evidence is missing, the agent must trigger a HALT.
+
+## Core Rule 3 — No architecture before KPI/source validation
+
+The agent must not create final architecture, storage, ingestion, transformation, or serving designs until:
+- primary KPI feasibility is verified or explicitly caveated
+- P1 sources are access-verified
+- source sample/schema evidence exists
+- grain is approved by the user
+- trust expectation is approved by the user
+- freshness expectation is approved by the user
+
+If these are missing, trigger:
+
+`HALT-10 — Architecture Before Validation`
+
+## Core Rule 4 — Decision gates are mandatory learning points
+
+Every critical decision must pass through a Decision Gate.
+
+At each Decision Gate, the agent must:
+- explain why the decision matters
+- connect the decision to the data engineering lifecycle
+- show current evidence
+- present options
+- recommend one option
+- explain trade-offs
+- ask the user to choose
+- wait for explicit confirmation
+- record the final decision
+
+## Core Rule 5 — Preserve uncertainty
+
+If something is unknown, the agent must label it as:
+- Unknown
+- Assumption
+- Needs validation
+- Deferred
+- Rejected
+- Use with caveats
+
+The agent must not convert assumptions into facts.

package/src/dew-skills/1-analysis/dew-decision-gate/checklist.md

@@ -0,0 +1,31 @@
+# Decision Gate Checklist
+
+Use this checklist before closing any DEW decision gate.
+
+## Required
+
+- [ ] Gate ID is identified.
+- [ ] Phase is identified.
+- [ ] Trigger is explained.
+- [ ] Concept is explained.
+- [ ] Current evidence is shown.
+- [ ] Missing evidence is shown.
+- [ ] Options are provided.
+- [ ] Recommendation is provided.
+- [ ] Trade-offs are explained.
+- [ ] User explicitly chooses an option.
+- [ ] `.decision-log.md` is updated.
+- [ ] `.learning-log.md` is updated when learning mode is enabled.
+- [ ] Artifact frontmatter is updated when applicable.
+- [ ] Evidence index is updated when applicable.
+- [ ] Resume condition is satisfied.
+
+## Failure conditions
+
+Fail the checklist if:
+- agent chose silently
+- no user confirmation exists
+- no decision log entry exists
+- missing evidence is hidden
+- assumptions are presented as facts
+- next phase begins before required gate is resolved

package/src/dew-skills/1-analysis/dew-decision-gate/customize.toml

@@ -0,0 +1,30 @@
+# DO NOT EDIT -- overwritten on every update.
+#
+# Workflow customization surface for dew-decision-gate.
+#
+# Override files:
+#   {project-root}/_dew/custom/dew-decision-gate.toml
+#   {project-root}/_dew/custom/dew-decision-gate.user.toml
+
+[workflow]
+
+activation_steps_prepend = []
+activation_steps_append = []
+
+persistent_facts = [
+  "file:{project-root}/**/project-context.md",
+  "file:{project-root}/**/.decision-log.md",
+  "file:{project-root}/**/.learning-log.md",
+]
+
+on_complete = ""
+
+global_rules = "assets/global-rules.md"
+gate_template = "assets/decision-gate-template.md"
+gate_catalog = "assets/gate-catalog.md"
+frontmatter_schema = "assets/gate-frontmatter-schema.md"
+checklist = "checklist.md"
+
+decision_log_path = "{planning_artifacts}/.decision-log.md"
+learning_log_path = "{learning_artifacts}/.learning-log.md"
+evidence_index_path = "{evidence_artifacts}/validation-evidence-index.md"

package/src/dew-skills/1-analysis/dew-evidence-check/SKILL.md

@@ -0,0 +1,104 @@
+---
+name: dew-evidence-check
+description: Check whether a DEW phase has the validation evidence required before it can be marked complete.
+---
+
+# DEW Evidence Check
+
+**Goal:** Prevent artifact-only completion.
+
+**Your Role:** You are an evidence auditor.
+
+You verify whether a phase can be marked done.
+
+## Conventions
+
+- Bare paths resolve from the skill root.
+- `{skill-root}` resolves to this skill's installed directory.
+- `{project-root}`-prefixed paths resolve from the project working directory.
+- `{workflow.<name>}` resolves to fields in `customize.toml`'s `[workflow]` table.
+
+## Mandatory Rules
+
+- Planning artifacts are not evidence.
+- A phase is not done without evidence.
+- Missing evidence must trigger HALT-18.
+- Caveated continuation requires explicit user approval.
+
+## On Activation
+
+1. Resolve customization:
+
+   `python3 {project-root}/_dew/scripts/resolve_customization.py --skill {skill-root} --key workflow`
+
+   If the script fails, read `{skill-root}/customize.toml` directly and use defaults.
+
+2. Execute `{workflow.activation_steps_prepend}`.
+
+3. Load persistent facts from `{workflow.persistent_facts}`.
+
+4. Load config from `{project-root}/_dew/dew/config.yaml` if present.
+
+5. Load:
+   - `{workflow.phase_done_checklist}`
+   - `{workflow.evidence_index_template}`
+
+6. Greet user in configured language.
+
+## Execution
+
+### Step 1 — Identify phase
+
+Identify:
+- phase name
+- target artifact
+- required gates
+- required evidence
+
+### Step 2 — Inspect evidence
+
+Check:
+- validation evidence index
+- evidence folder
+- artifact frontmatter
+- decision log
+- learning log
+
+### Step 3 — Classify evidence
+
+Classify each required evidence item as:
+- missing
+- available
+- validated
+- stale
+- invalid
+
+### Step 4 — Produce evidence report
+
+Output:
+
+| Requirement | Status | Evidence Path | Notes |
+|---|---|---|---|
+
+### Step 5 — Decision
+
+If all required evidence is valid:
+- phase can be marked done
+
+If evidence is missing:
+- trigger HALT-18
+
+If evidence is partial:
+- ask user whether to:
+  A. produce missing evidence
+  B. continue with caveat
+  C. reduce scope
+  D. return to previous phase
+
+### Step 6 — Record
+
+Update:
+- `.decision-log.md`
+- `.learning-log.md`
+- validation evidence index
+- artifact frontmatter