authorship-strategy-skill
shimo4228/authorship-strategy-skill
Claude Code skill: operational form of the authorship-strategy four-layer judgment framework (three-axis inversion, four-layer stack, prohibited and encouraged actions, 11-item judgment checklist). Loadable into LLM-based coding agents as a rule set.
Instalação
npx convez-ai shimo4228/authorship-strategy-skill1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
README
A tradução para o português deste README ainda está na fila.
authorship-strategy-skill
A Claude Code skill that loads the four-layer judgment framework for authorship strategy under AI-mediated diffusion into an LLM-based coding agent's rule set. Provides the operational form (trigger conditions, judgment checklist, prohibited and encouraged actions) of the normative framework articulated in the authorship-strategy research line.
The skill is scoped to a narrow trigger: it applies only to the author's own DOI-targeted idea-rescue research repositories, not to general client work or commercial deliverables. Adopting the skill commits the agent to the framework's core principle that authorship strategy under AI-mediated diffusion inverts on three axes (scarcity → diffusion, exclusivity → derivation, enclosure → openness).
When to use
Apply the skill when all of the following hold:
- The current repository owner is the skill's adopter (the framework is single-author normative; adapting it to client work or contributions to others' projects is out of scope)
- The repository is a DOI-targeted or DOI-registered research artifact (Zenodo or comparable archive registration intended or in place)
- The artifact is an idea-rescue in character — specifications, schemas, ADRs, glossaries, or doctrine documents whose value is at the concept level rather than at the implementation level
Skip the skill if any one of those is missing. Specifically, skip for:
- Client work and commercial deliverables (monetization is the goal; the author's identity is not the protected value)
- Contributions to others' projects (the other author's strategy takes precedence)
- Operational tooling and scaffolding meant to dissolve (the artifact's diffusion outside its substrate is not desired)
- Day-to-day coding and debugging (the framework would be over-applied as a judgment instrument)
When in doubt, the framework defers to the author's explicit judgment rather than auto-applying.
Install
Claude Code
git clone https://github.com/shimo4228/authorship-strategy-skill
cp -r authorship-strategy-skill/skills/authorship-strategy ~/.claude/skills/authorship-strategy
No runtime dependencies. The skill is documentation-only; it shapes agent judgment by being loaded into the agent's rule set.
Other harnesses
The skill body is a standard markdown file with a YAML frontmatter header. Adapt the install path to your harness's skill convention.
How it works
- Trigger gate — the skill checks four conditions (DOI-targeted, idea-rescue character, author's own work, not commercial) before applying. Misfires are explicitly cheaper than false negatives; when in doubt, the skill defers to author confirmation.
- Three-axis inversion as the core principle — value source (scarcity → diffusion), validation mechanism (exclusivity → derivation), network effect (enclosure → openness). The three axes co-vary; mixing axes is internally inconsistent.
- Four-layer judgment stack — Authenticity (Layer 1, the value) → Attribution Diffusion (Layer 2, the strategy) → Idea vs Scaffold (Layer 3, the prediction) → Tactics (Layer 4, the decisions). Each layer is downstream of the layer above; tactical decisions justified at Layer 4 must trace back to Layer 1.
- Prohibited and encouraged actions — the skill blocks proposing monetization strategies, exclusionary competitive positioning, market-driven concept dilution, sensational framing, and over-scoped origin claims. It encourages DOI-citable structures, distinctive terminology coinage, abstract-doctrine-plus-worked-implementation pairs, friction-minimized adoption paths, and permissive licensing.
- Judgment checklist — fifteen yes/no questions the agent runs against every proposal in the framework's trigger scope.
- Operating the strategy over time — a two-tier discipline for running the strategy across sessions: a private implementation ledger (operational status, ranked candidate interventions) versus a public, dated, effect-claim-free intervention timeline, plus a periodic gap-review (deployed tactics against the tactic catalog and open questions) that generates the next proposals.
Key concept: three-axis inversion is structural, not stylistic
The framework's strongest claim is that the three axes are not independent dimensions an author can freely mix. A strategy combining enclosure with derivation-welcoming license, or scarcity with crawler-friendly access, is internally inconsistent and weaker than either fully consistent strategy. The skill enforces this consistency: a Layer 4 tactic that violates the inversion on any axis is flagged regardless of how locally attractive it appears.
What this skill does NOT do
| Concern | Use this instead |
|---|---|
| Release-time workflow for DOI-registered repos (verify, tag, deposit, propagate DOI) | release-doi |
| llms.txt / llms-full.txt prose design, navigator wording | llms-txt-writer |
| JSON-LD knowledge graph design | jsonld-knowledge-graph |
| Cross-document drift audits | context-sync |
| Article / blog post writing | claude-skill-writing-ecosystem |
Related research and skills
- Doctrine repository: authorship-strategy — the normative framework, five tactical ADRs, and empirical baseline this skill is the operational form of
- Peer components (other component skills of the same framework):
- release-doi — release-time workflow operationalizing the identifier-federation triplet (ADRs 0001-0003)
- llms-txt-writer — operationalizes Layer 4 tactic 7's Answer.AI
llms.txtconvention - jsonld-knowledge-graph — operationalizes Layer 4 tactic 7's JSON-LD knowledge graph
- Sibling research lines (at the research-program level): Agent Knowledge Cycle (AKC), Contemplative Agent, Agent Attribution Practice (AAP)
Terminology note. This ecosystem reserves sibling for research-line-level peers; at the component-skill level the term peer component is used instead.
License
MIT. See LICENSE.