edgy-deep-dive
Targeted EDGY 23 deep-dive analysis of a specific element pair or facet combination. Requires an existing edgy-model.json produced by the edgy-assessment skill. Analyses the chosen elements through one of four analytical lenses (dependency, alignment, gaps, opportunities) using both direct core links and 2-hop paths. Produces a focused markdown report and a pairwise draw.io diagram.
- License 许可
- Apache-2.0
- 最近更新
- 2026-10-19
- Stars
- ★ 2
- Forks
- 0
01 前言 Front Matter
- name
- edgy-deep-dive
- tags
- [ "edgy", "enterprise-design", "architecture", "analysis", "deep-dive", "intersection", "fi", "en" ]
- agents
- [ "claude-code", "cursor", "vibe", "generic" ]
- inputs
- [ { "name": "model", "type": "file", "description": "edgy-model.json produced by the edgy-assessment skill. Contains all 12 EDGY elements, active core links, coherence scores, and suggested deep-dive analyses for this company.\n" }, { "name": "focus_elements", "type": "array", "description": "2–3 EDGY element names to analyse. Use official names: purpose, story, content, capability, asset, process, task, channel, journey, organisation, product, brand. Examples: [\"capability\", \"organisation\"] or [\"task\", \"channel\", \"journey\"]\n" }, { "name": "lens", "type": "enum", "values": [ "dependency", "alignment", "gaps", "opportunities", "all" ], "default": "all", "description": "Analytical lens to apply. dependency — how the elements need each other; critical dependencies alignment — coherence between elements; contradictions gaps — missing links or underdeveloped elements opportunities — competitive advantage or growth potential all — all four lenses\n" }, { "name": "language", "type": "enum", "values": [ "fi", "en" ], "default": "fi", "description": "Output language. Currently only fi and en are supported (output templates and lens-specific section headers are localised for these two only).\n" } ]
- outputs
- [ { "type": "markdown", "description": "Focused deep-dive analysis with direct and 2-hop core-link analysis" }, { "type": "file", "description": "edgy-diagram-compatible TXT input for pairwise map" }, { "type": "file", "description": "draw.io diagram (pairwise layout, focus elements bold-bordered)" } ]
- version
- 1.1.2
- category
- architecture
- languages
- [ "fi", "en" ]
- description
- Targeted EDGY 23 deep-dive analysis of a specific element pair or facet combination. Requires an existing edgy-model.json produced by the edgy-assessment skill. Analyses the chosen elements through one of four analytical lenses (dependency, alignment, gaps, opportunities) using both direct core links and 2-hop paths. Produces a focused markdown report and a pairwise draw.io diagram.
02 资源文件 2
- 参考文档examples/expected-deep-dive.md4.5 KB
- 参考文档examples/sample-model.json9.3 KB
03 SKILL.md 正文
EDGY Deep-Dive Analysis
Purpose
This skill performs a targeted analysis of a specific EDGY element pair using an existing company EDGY model as context. It traces both direct core links and indirect (2-hop) paths between the chosen elements, then analyses their relationship through the selected lens.
Use this skill when:
- You have completed an
edgy-assessmentand want to explore a specific element pair in depth - The assessment's "Suggested deep-dive analyses" section identified a High-priority pair
- The client wants to understand how, for example, Architecture capabilities are structured relative to the Organisation, or how Products map to customer Tasks
Prerequisite: edgy-assessment must have been run and produced <company>-edgy-model.json.
This skill produces:
<company>-deep-dive-<elements>.md— focused analysis report<company>-deep-dive-<elements>.txt— edgy-diagram input (pairwise map)<company>-deep-dive-<elements>.drawio— pairwise draw.io diagram
IMPORTANT: Produce all output in the language specified by the language parameter.
Official 24 EDGY Core Links (reference for path tracing)
<!-- edgy-links:begin format=flat-fi-en -->
| Source → Target | FI | EN |
|---|---|---|
| story → purpose | kontekstualisoi | contextualises |
| content → purpose | ilmaisee | expresses |
| content → story | välittää | conveys |
| brand → story | herättää | evokes |
| brand → purpose | edustaa | represents |
| capability → asset | vaatii | requires |
| process → capability | toteuttaa | realises |
| process → asset | vaatii | requires |
| product → capability | vaatii | requires |
| process → product | luo | creates |
| task → journey | on osa | is part of |
| task → channel | käyttää | uses |
| journey → channel | kulkee | traverses |
| brand → task | tukee | supports |
| brand → journey | näkyy | appears in |
| organisation → purpose | tavoittelee | pursues |
| organisation → story | kirjoittaa | authors |
| organisation → capability | omistaa | has |
| organisation → process | suorittaa | performs |
| product → task | palvelee | serves |
| product → journey | esiintyy | features in |
| organisation → brand | rakentaa | builds |
| organisation → product | valmistaa | makes |
| product → brand | ilmentää | embodies |
| <!-- edgy-links:end --> |
Agent Instructions
CRITICAL: Read the full instructions before starting
Do not skip phases. Every phase is required to produce a complete, grounded analysis.
Phase 1: Load Context
- Validate the model against the shared schema before reading it:
python3 <edgy-assessment>/scripts/validate_edgy_model.py <company>-edgy-model.json(schema:skills/architecture/edgy-assessment/assets/edgy-model.schema.json). A model that fails validation is not analysed — ask the user to fix it or to runedgy-assessmentinextract-modelmode. Never invent elements that the model does not contain. - Read the
modelfile (edgy-model.json).elements.organisationis one object (format 1.8) or an array (1.9), and so iselements.journeyfrom format 1.10 (one object, or an array of journeys with ids) — treat a single object as a list of one. With several journeys, analyse the ones the focus names, or each in turn; never merge them into one. - Extract the entries for all
focus_elementsfromelements. - Note the coherence scores from
coherencefor any intersection elements in the focus. - Check
suggested_deep_dives— if this exact pair was suggested, note its rationale and priority.
Minimum requirement: You must extract at least one element description and at least one active core link from the model before proceeding. If the model is missing or malformed, ask the user to run edgy-assessment first.
Phase 2: Trace Core-Link Paths
Build the link graph for the focus elements:
Step 1 — Direct links (1-hop)
List every core link from the table above where both source and target are in focus_elements.
Example for ["capability", "organisation"]:
organisation → capability: has✓ (direct)
Step 2 — 2-hop paths
List every path A → X → B (or B → X → A) where A and B are in focus_elements and X is any other element.
Example for ["capability", "organisation"]:
organisation → process: performsANDprocess → capability: realises→ path: organisation →(performs)→ process →(realises)→ capability
Step 3 — Cross-check with active model links
From the model's core_links array, mark which of the above links are active (present in the model) vs. theoretical (in the framework but not observed in this company). An absent link that should exist is itself a gap finding.
Present results as a table:
| Path | Hops | Status | Verb(s) |
|---|---|---|---|
| organisation → capability | 1 | Active | omistaa |
| organisation → process → capability | 2 | Partial (process→capability missing) | suorittaa / toteuttaa |
Phase 3: Deep Analysis by Lens
Run the analysis for each selected lens. Every finding must reference a specific element from the model (use element names, not generic statements).
Lens: Dependency (dependency)
Analyse how the focus elements structurally need each other:
- Critical dependencies: Which links are load-bearing? What breaks if the link is absent?
- Direction of dependency: Which element depends on the other, or is it mutual?
- Bottlenecks: If one element changes (e.g. a new capability is added), which element must change too?
- Resilience: Are there alternative paths that reduce single-point dependency?
Output format:
### Riippuvuusanalyysi: [Elementti A] ↔ [Elementti B]
**Kriittiset riippuvuudet:**
- [elementti] vaatii [toinen elementti] koska ... (lähde: malli-elementti)
**Riippuvuuden suunta:** [yksisuuntainen A→B | molemminpuolinen]
**Pullonkaulat:**
- ...
**Vaihtoehtoiset polut:**
- ...
Lens: Alignment (alignment)
Analyse coherence and contradictions between the focus elements:
- Alignment score per link: Strong / Good / Weak — with justification
- Contradictions: Explicit conflicts between element descriptions or their relationship
- Mismatches: Areas where element A implies something that element B doesn't support
- Root cause hypothesis: Why does the misalignment exist?
Output format:
### Yhdenmukaisuusarvio: [Elementti A] ↔ [Elementti B]
| Linkki | Yhdenmukaisuus | Perustelu |
|--------|----------------|-----------|
| organisation → capability | Weak | Org-rakenne on toiminnallinen, kyvykkyydet ovat asiakkuuskohtaisia |
**Ristiriidat:**
- ...
**Juurisyyanalyysi:**
- ...
Lens: Gaps (gaps)
Identify missing links and underdeveloped elements:
- Missing core links: Links that should exist (per framework) but are absent from the model
- Underdeveloped elements: Elements with only superficial descriptions or no active links
- Consequence of each gap: What business problem does this gap cause?
- Gap severity: High (blocks strategic objectives) / Medium (slows delivery) / Low (nice to have)
Output format:
### Puuteanalyysi: [Elementti A] ↔ [Elementti B]
| Puuttuva linkki | Pitäisi olla | Vaikutus | Vakavuus |
|----------------|-------------|---------|---------|
| process → capability | process toteuttaa capability | Prosessit eivät tue strategisia kyvykkyyksiä | High |
**Alikehittyneet elementit:**
- [elementti]: vain yksi kuvaus, ei aktiivisia linkkejä → ...
Lens: Opportunities (opportunities)
Identify where the element pair creates competitive advantage or growth potential:
- Leverage points: Where is the element pair unusually strong vs. industry baseline?
- Compounding effects: Where could investing in one element multiply value in the other?
- Strategic asymmetry: What can this company do with this element combination that competitors cannot?
- Priority actions: Top 3 actions to realise the opportunity, with responsible element
Output format:
### Mahdollisuudet: [Elementti A] ↔ [Elementti B]
**Vipuvoimapisteet:**
- ...
**Kerrannaishyödyt:**
- [elementti A]:n vahvistaminen → vaikuttaa [elementti B]:hen koska ...
**Strateginen epäsymmetria:**
- ...
**Prioriteettiset toimenpiteet:**
1. [toimenpide] — vastuullinen elementti: [X] — vaikutus: High/Medium/Low
2. ...
3. ...
Phase 4: Generate Pairwise Diagram Input
Produce <company>-deep-dive-<elements>.txt in edgy-diagram input format.
Include only:
- The focus elements themselves
- Their direct (1-hop) neighbours that participate in active links
- All core links between these elements
Mark focus elements with [focus] tag.
The TXT file is a human-readable specification — the edgy_parser.py does
not currently parse pairwise layouts, so the agent generates the .drawio
XML directly following the pairwise layout spec in
skills/documentation/edgy-diagram/SKILL.md (see "Pairwise Map" section).
Use the same facet / elements / relationships structure as other
edgy-diagram inputs for readability:
facet: architecture
focus: [capability, organisation]
elements:
- capability: "Capability name - Description" [focus]
- organisation: "Organisation name - Description" [focus]
- process: "Process name - Description" (neighbour via organisation→process→capability)
relationships:
- "organisation" -> "capability": "omistaa"
- "organisation" -> "process": "suorittaa"
- "process" -> "capability": "toteuttaa"
Then author <company>-deep-dive-<elements>.drawio directly as XML following the pairwise layout rules (the generator does not yet support the pairwise layout), include the legend block from the edgy-diagram skill, and lint the file before delivery:
G=${EDGY_DIAGRAM_SCRIPTS:-skills/documentation/edgy-diagram/scripts} # installed skills: ~/.claude/skills/edgy-diagram/scripts
python3 $G/edgy_lint.py <company>-deep-dive-<elements>.drawio # 0 errors required
python3 $G/edgy_render.py <company>-deep-dive-<elements>.drawio # SVG (+PNG) preview — look at it before delivery
- Focus elements:
strokeWidth=4(bold border), same colour as their element type - Neighbour elements: normal
strokeWidth=2 - Direct core links: solid arrow + verb label
- 2-hop paths: dashed arrow (
dashed=1) + intermediate element name as path label - Layout: focus element A left column, focus element B right column, neighbours outside both
Phase 5: Deliver and Suggest Next Step
- Show the complete
<company>-deep-dive-<elements>.mdreport. - End with a "Seuraava jatkoanalyysi" (Next deep-dive) suggestion block:
## Seuraava jatkoanalyysi
Tämän analyysin löydösten perusteella suositellaan seuraavaa:
| Elementtipari | Linssi | Perustelu |
|---------------|--------|-----------|
| capability × process | gaps | [Puute löydettiin: process→capability-linkki puuttuu] |
| product × task | alignment | [Löydettiin tuote jolla ei ole task-linkkiä] |
The suggestion must be grounded in a specific finding from this analysis (reference phase 3 outputs).
Output Template
Finnish (fi)
# EDGY Syväanalyysi: [Elementti A] × [Elementti B]
**Yritys:** [Nimi]
**Linssi:** [dependency | alignment | gaps | opportunities | all]
**Analysoitu:** [pvm]
---
## Linkkartta
[Phase 2 table — direct links and 2-hop paths]
---
## [Linssi-osio(t)]
[Phase 3 content]
---
## Yhteenveto
[2–3 lausetta päälöydöksistä ja suositellusta toimenpiteestä]
---
## Seuraava jatkoanalyysi
[Phase 5 suggestion table]
English (en)
# EDGY Deep-Dive: [Element A] × [Element B]
**Company:** [Name]
**Lens:** [dependency | alignment | gaps | opportunities | all]
**Analysed:** [date]
---
## Link Map
[Phase 2 table]
---
## [Lens section(s)]
[Phase 3 content]
---
## Summary
[2–3 sentences on key findings and recommended action]
---
## Next Deep-Dive
[Phase 5 suggestion table]
Quality Gate
Before delivering, verify:
- [ ] All focus elements are present in the model (not invented)
- [ ] Link map contains at least 1 direct or 2-hop link
- [ ] Every finding references a specific element name from the model
- [ ] No generic statements ("the organisation should improve its capabilities") without specific grounding
- [ ] Pairwise diagram TXT contains
[focus]tags on focus elements - [ ]
edgy_lint.pyreports 0 errors for the pairwise drawio file - [ ] The pairwise diagram was previewed with
edgy_render.pyand the focus/neighbour layout reads correctly - [ ] Next deep-dive suggestion is grounded in a finding from this analysis
Anti-patterns
- DO NOT analyse elements that are not in the model without noting they are missing
- DO NOT write generic capability/organisation advice not grounded in model data
- DO NOT skip the link map — it is the structural foundation of the analysis
- DO NOT suggest the same element pair as the next deep-dive (suggest a new pair)
Reference
See examples/ for a complete deep-dive example:
examples/sample-model.json— Globex Oy EDGY model (from edgy-assessment)examples/expected-deep-dive.md— capability × organisation, lens=dependency