skills.rest
alvis
alvisCOMMUNITY
Explore all skills by this author →
Website𝕏@alvishttangGitHub

discover

Explores unknowns before planning and records findings in an evidence ledger.

4|1|Updated Jul 31, 2025
Search Tags:#planning#brainstorming#discovery#blindspots#readiness check#evidence ledger#interactive boards
One-click install
npx skills add https://github.com/alvis/.agents --skill discover-alvis
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: discover
Source: https://github.com/alvis/.agents/tree/main/plugins/essential/skills/discover
Command: npx skills add https://github.com/alvis/.agents --skill discover-alvis

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? It reduces consequential uncertainty before implementation by surfacing blindspots, latent preferences, and unverified assumptions, so plans are not built on guesses that later prove false. ## Core Features & Use Cases - Eight discovery modes: blindspots, options, interview, reference, prototype, readiness, state, and implementation, each with its own direction file and workflow. - Evidence ledger: every finding is classified as intent, observed, inference, unknown, or assumption, with provenance, decision impact, reversibility, and an owner. - Interactive HTML boards: renders self-contained, annotatable review pages from JSON data via the bundled renderer, capturing user answers into a single generated prompt. - Use Case: Before planning a migration, run a blindspot pass to surface hidden constraints, then present ranked options on an interactive board and hand the confirmed evidence ledger to the decision skill. ## Quick Start Ask the assistant to run a blindspot discovery pass on your problem, for example: discover the unknown unknowns in migrating our billing service to event-driven architecture.

Frequently Asked Questions about discover

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I find unknown unknowns before starting a project?▼

Run the blindspots mode, which buckets known knowns, known unknowns, latent preferences, and blindspot hypotheses, then inspects the smallest relevant code or integration surface to test each hypothesis. Findings are ranked by decision impact and reversibility.

What is the difference between discover and decide skills?▼

Discover owns divergent exploration: surfacing options, unknowns, and evidence. Decide owns converging on one approach among already-grounded options. Discover passes its evidence ledger to decide when the option set is ready.

Can discovery results be presented as interactive HTML pages?▼

Yes. The skill renders self-contained boards from JSON data using scripts/render-page/cli.ts, with annotation, per-card response capture, and a single generated reply prompt. Boards are temporary review surfaces discarded after decisions transfer to the ledger.

When should I not use a discovery skill?▼

Skip it for fact-finding reports, metric-driven optimization, choosing among already-grounded options, production UI design, or clear bounded implementation work. Those route to deep-research, autoresearch, decide, design, or implementing skills respectively.

How are discovery findings persisted across sessions?▼

With --persist, the ledger is written to state/discovery.md as a child of a registered DSC parent task in the root state file, with journal entries per material finding. Disposable prototypes live under artifacts/prototypes/ and generated HTML stays in an OS temporary directory.

Related Skills

Explore More Software Engineering Skills→
anthropicsanthropics
23.4k

write-spec

Generate structured PRDs and feature specifications from problem statements.

Official
Advanced
linenoizelinenoize
1

brainstorm

Compare solution options with trade-offs and recommend one path.

Community
Advanced
Tyler-R-KendrickTyler-R-Kendrick
1

autoresearch

Coordinates research loops linking training pipelines with curated documentation and Linear tracking.

Community
Advanced
zivtechzivtech

jhodgdon-compare

Compare Drupal contributor positions using public issue queue evidence with CID citations.

Official
Advanced