<< All versions
Skill v1.0.0
currentAutomated scan100/100zenx0x/allinluna/research-routes-explore
──Details
PublishedNovember 4, 2026 at 06:58 PM
Content Hashsha256:ec0454c9b80b693c...
Git SHAb471713089d8
──Files
Files (1 file, 1.8 KB)
SKILL.md1.8 KBactive
SKILL.md · 23 lines · 1.8 KB
name: research-routes-explore description: Explore open-ended scientific or product questions through multiple concurrent Research Routes branches, probes, experiments, observations, evidence polarity, failure regimes, unknowns, and mature-method comparators. Use for research exploration or hybrid discovery from any starting point while keeping a route-neutral terrain map separate from claims, authorization, implementation, and canonical state. metadata: version: "1.0.0"
Research Routes Explore
Explore without collapsing uncertainty. The Research Pack is append-only and route-neutral.
- Start from the supplied question, problem, frame, repository, artifact, or arbitrary context; record it as the starting point.
- Build a terrain map and shared backbone before selecting a route.
- Keep at least two viable routes or explicitly record why only one route is available. Branch routes rather than overwriting them.
- Design probes and experiments with boundary conditions and failure regimes. Record observations as evidence with polarity and provenance.
- Preserve contradictions and unknowns. A null or failure result is still an observation and does not silently invalidate a route.
- Keep candidate inference, experiment authorization, HumanDecision, implementation, and canonical promotion as separate transitions.
- Compare mature methods and alternative explanations when the question is scientific; preserve their scope and limitations.
- Record failures with polarity and what did not fail, then use rewind, lessons, and reopened problems to recover without rewriting history.
- Do not treat a terrain map as a route choice, experiment authorization, implementation, HumanDecision, or canonical-state promotion.
The shared semantic model and schema are authoritative for node, relation, polarity, route, and lifecycle values.