
from DeepScientist2,620
Frame ambiguous research or engineering tasks: clarify goals, metrics, datasets, and a justified baseline so work can proceed with confidence.
Scout formalizes early-stage research framing. When a task lacks a clear objective, metric, dataset contract, or baseline, this skill runs a focused, minimal discovery pass to produce a durable next anchor (e.g., a baseline shortlist or a clarified evaluation contract). The goal is not exhaustive literature review but to eliminate the framing ambiguities that block progress.
Use Scout when the user's goal is still ambiguous, the dataset or split contract is unclear, the primary metric is undefined, no trustworthy baseline exists, or the project has been paused and needs framing reconstruction. Avoid Scout when a validated baseline and clear execution plan already exist.
references/* contains paper-triage, literature-scout templates, and evaluation-contract templates used to structure the pass.Primarily useful for research-focused agents and workflows (agents with web/paper search, memory, and artifact access). Works well when the agent has access to memory.search, artifact.arxiv, and repository artifacts for durable context.
DeepScientist Scout is a purely instructional skill (no bundled scripts) for framing ambiguous research tasks within the DeepScientist framework. The SKILL.md is well-structured with clear match signals, pitfalls, constraints, validation checks, and exit criteria. It defines a 5-step control workflow for resolving minimum framing unknowns before handing off to baseline or idea stages. No scripts to execute — static analysis only.
No security concerns — no scripts, no network calls, no credentials. Clean instructional skill. Niche usefulness due to DeepScientist platform dependency.