Skill
Effects Architecture Skill
Generate verified architecture documentation using the effectHandler pattern.
From devacInstall
1
Run in your terminal$
npx claudepluginhub pietgk/vivief --plugin devacTool Access
This skill uses the workspace's default tool permissions.
Supporting Assets
View in Repositoryknowledge/quality-rules.mdknowledge/tacit-insights.mdSkill Content
Effects Architecture Skill
Generate verified architecture documentation using the effectHandler pattern.
Related Skills
- validate-architecture: Uses M1-M4 rules for C4 architecture validation. Also defines G1-G4 gap metrics specific to C4 quality measurement. See @plugins/devac/skills/validate-architecture/SKILL.md
Foundational Principle
effectHandler = (State, Effect) => (State', [Effects'])
Verified sources:
- @docs/vision/concepts.md:15 - Core definition
- @docs/vision/foundation.md:22 - Equivalence principle
- @docs/adr/0015-conceptual-foundation.md:24-28 - ADR formalization
Triggers
This skill activates when users ask about:
- "explain effects", "how do effects work"
- "generate architecture doc", "document effects architecture"
- "verify this documentation"
- "what is an effectHandler"
Capabilities
1. Answer Effect Questions
Use conceptual framework + DevAC tools to answer accurately. Always verify claims against source documents.
2. Generate Verified Documentation
Create documents like how-are-effects-handled-v4.md with:
- Confidence markers (HIGH/MEDIUM/LOW/N/A)
- Source verification (file:line references)
- Complete enumerations (apply M1)
- Generic-first explanations (apply M2)
3. Apply Quality Rules
Before finalizing any output, check against M1-M4 rules. See @knowledge/quality-rules.md
4. Use Tacit Insights
Incorporate knowledge not detectable from code/DevAC. See @knowledge/tacit-insights.md
Knowledge Sources
| Source | Purpose |
|---|---|
| @knowledge/tacit-insights.md | Insights not in code/docs |
| @knowledge/quality-rules.md | M1-M4 meta-rules |
| @examples/effects/how-are-effects-handled-v4.md | Reference output (clean, M1-M4 compliant) |
| @examples/effects/how-are-effects-handled-v3.md | Historical version (has known issues) |
| @examples/effects/prompts.md | Prompt evolution |
| @examples/effects/v3-mistakes-analysis.md | Lessons learned |
| @docs/vision/concepts.md | Conceptual foundation |
Workflow
For Questions
- Read tacit-insights.md for knowledge not in DevAC
- Query DevAC tools for code-level data
- Verify against source documentation with file:line
- Answer with confidence markers
For Document Generation
- Read all knowledge sources
- Query DevAC for effects, nodes, edges, schema
- Apply M1 before any enumeration (check set completeness)
- Apply M2 for any concept explanation (generic -> varieties -> example)
- Add confidence markers to all claims
- Include Data Sources chapter with verification commands
CLI Commands
# Query effects
devac effects list -p <package>
# Query schema
devac query "SELECT * FROM effects LIMIT 5"
# Get file symbols
devac file-symbols <path>
MCP Tools
query_effects- Query raw effectsget_schema- Get table structurerun_rules- Execute handler dispatchlist_rules- Show available rules
Example Interaction
User: "How do effects work in DevAC?"
Response approach:
- Read @knowledge/tacit-insights.md for Parser-as-Handler, AST-as-State
- Start with generic definition from concepts.md:15
- Show implementation varieties (M2): pattern matching, AST traversal, direct transformation
- Give specific example: RuleEngine
- Mark confidence levels
- Cite sources with file:line
User: "Generate architecture documentation for effects"
Response approach:
- Use @examples/effects/how-are-effects-handled-v4.md as template
- Query DevAC for current data
- Apply M1: count sets before listing (4 parquet files, list all 4)
- Apply M2: define concepts generically before examples
- Include Data Sources chapter
- Add verification commands
Notes
- Always read tacit-insights.md before answering effect questions
- Apply M1-M4 rules to all generated documentation
- Reference existing example docs rather than duplicating
- Mark confidence: HIGH (source code), MEDIUM (MCP), LOW (inferred), N/A (planned)
Similar Skills
Stats
Parent Repo Stars4
Parent Repo Forks0
Last CommitJan 11, 2026