| name | structural-issue-authoring |
| description | Portable skill for authoring GitHub issues as structural tension charts. Five questions → five sections. Works in any repository, any codebase. Agents install this skill to shift issue authoring from reactive bug-report framing to generative creative orientation. No repo-specific dependencies. Grounded in Robert Fritz's structural dynamics and creative orientation.
|
Structural Issue Authoring
Forge issues as structural tension charts — in any repo, any codebase.
This is a portable skill. It has no dependency on any specific repository structure. Any agent that reads this file can author issues using structural tension.
The Five Questions
Before writing an issue, answer five questions. Each produces one section:
| # | Question | Section | What it captures |
|---|
| 1 | What exists now? | Context | Current reality — factual, specific |
| 2 | What do I want to exist? | Desired State | The creation, not the problem |
| 3 | What moves me from one to the other? | Action Steps | Strategic secondary choices |
| 4 | What is the tension? | Structural Tension | The causal force made visible |
| 5 | What is this related to? | Related | Kinship — nothing exists alone |
Template
## Context
[Current reality — what exists now. Factual, specific, no motivation language.
No "ready to begin." No "excited to start." Just what is.]
## Desired State
[The creation you want to bring into being.
Describe the result, not the process.
"Auth flow that honors session lifecycle" not "Fix broken auth."]
## Action Steps
- [ ] [Strategic secondary choice 1]
- [ ] [Strategic secondary choice 2]
- [ ] [...]
## Structural Tension
**Current reality:** [one-line restatement]
**Desired state:** [one-line restatement]
## Related
- [issues, repos, ceremonies, upstream/downstream — kinship connections]
Authoring Principles
1. Create, don't fix
The issue describes something being brought into existence, not a problem being eliminated. This is the foundational shift.
| ❌ Reactive (problem-solving) | ✅ Generative (creative orientation) |
|---|
| "Fix broken auth flow" | "Auth flow that honors session lifecycle" |
| "Missing docs for X" | "Guidance for X that agents can act on" |
| "Bug: list renders wrong" | "List rendering that preserves structural intent" |
| "Refactor messy utils" | "Utils organized by domain responsibility" |
2. Context = Current Reality
- Objective and factual — what exists right now
- No motivation language — not "eager to," "ready to," "need to"
- Specific enough that someone else can verify whether it changed
- Reference commits, files, or observable state when possible
3. Desired State = Primary Choice
Follow Robert Fritz's desired outcome quality:
- Can you picture it? — Form a mental image of the result
- Quantify where possible — "5 platform guides" not "more documentation"
- Avoid comparative terms — "adequate test coverage" not "better tests"
- Describe results, not process — "API with typed endpoints" not "type the API"
- Be specific — "Auth middleware with JWT validation and refresh" not "improved auth"
4. Action Steps = Strategic Secondary Choices
Not a to-do list. Each step is chosen because it advances toward the desired state.
The test: "If we took these steps, would we achieve this result?"
If no → add more strategic steps or reconsider the desired state.
Each action step can telescope — become the desired state of its own issue. This is how large work decomposes naturally.
5. Structural Tension = The Engine
The two-line summary makes disequilibrium visible. Without it, the issue is a task list with no causal force. With it, readers understand why the work pulls forward.
The tension between current reality and desired state is not a gap to bridge — it is a force that resolves through creation.
6. Related = Kinship
Issues don't exist alone. Name:
- Upstream: what feeds into this work
- Downstream: what this work enables
- Siblings: parallel efforts in the same domain
- Ceremonies/contexts: the broader initiative this belongs to
When to Use
- Authoring any GitHub issue where the work is generative (building, designing, establishing)
- Converting observed patterns or session learnings into trackable work
- When an agent creates issues as part of its workflow
- Decomposing large desired outcomes into telescoped sub-issues
- Any context where reactive "fix/bug/broken" framing would obscure the creative intent
When NOT to Use
- Genuine incident reports where the reactive frame is accurate ("service down, restore it")
- Security advisories with prescribed remediation
- Simple chores with no creative tension ("bump dependency to v2.3")
Even in these cases, the five questions can clarify thinking — but forcing the template would be ceremony without substance.
Source
This skill is grounded in:
- Robert Fritz — structural tension, creative orientation, structural thinking
- Relational accountability — the Related section ensures nothing is authored in isolation
- Lived practice within the llms-txt ecosystem
Quick Reference Card
┌─────────────────────────────────────────┐
│ STRUCTURAL ISSUE AUTHORING │
│ │
│ 1. What exists now? → Context │
│ 2. What do I want? → Desired │
│ 3. What moves me there? → Steps │
│ 4. What's the tension? → Tension │
│ 5. What's related? → Related │
│ │
│ Create, don't fix. │
│ Tension resolves through creation. │
└─────────────────────────────────────────┘