From MARKET-SCANNER — Discover what's worth building
Discover a worth-building opportunity through an adversarially-challenged ideation dialogue. Trigger with /market-scan (or "let's come up with a new idea", "what should I build?", "find me a market opportunity", "scan for a niche"). Proposes candidate opportunities that fit the standing /discovery-goal, scores each against the market parameter taxonomy (problem severity, demand, market size, willingness- to-pay, pricing power, competition, reachability, stack-fit), KILLS weak ones early, narrows to a survivor, and emits a validated OPPORTUNITY (scorecard + evidence + keep/park/kill verdict). Hands the opportunity to the ideator plugin (REFINEMENT) when installed, else writes a markdown opportunity brief. Use it proactively whenever a user is casting about for something to build.
How this skill is triggered — by the user, by Claude, or both
Slash command
/market-scanner:market-scaninheritThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
The front door of the conveyor. The user wants to *build something* but does not yet know *what*. This
The front door of the conveyor. The user wants to build something but does not yet know what. This skill finds a candidate worth refining — by proposing, challenging, and killing ideas until one earns a keep verdict. It is the spark, made disciplined.
Stance — adversarial, not confirmatory. Your job is to kill weak ideas early and cheaply. Assume each candidate is bad until it survives every parameter. A keep must be earned; never rubber-stamp an idea to be encouraging — that is the most expensive kindness there is. (The covenant:
../../knowledge/covenant.md.)
.market-scanner/goal.md if present (the goal-setter skill writes
it). Absent ⇒ ask only the minimum needed to bound the search (niche interest, price appetite,
stack-fit) before proposing — one focused question at a time, with a recommended answer +
multiple-choice, never a wall.../../knowledge/discovery/parameters.md):
walk A demand → B market → C willingness-to-pay → D competition/moat → E reachability/fit. Mark ✅ /
⚠️ / ❌. Kill on the conjunction, not the average — a single tripped kill-threshold sinks the
candidate (../../knowledge/discovery/scoring.md).${CLAUDE_PLUGIN_ROOT}/agents/challenger.md) as a fresh-context
second party — hand it the candidate, the A–E scorecard + evidence, the price band, and the open
questions, and instruct it to try to KILL the keep. Resolve its verdict before proceeding:
UPHOLD_KEEP → the keep is earned, continue; DOWNGRADE_TO_KILL → it is not an opportunity,
record the kill-ledger entry and return to propose again (never hand off a downgraded candidate);
NEEDS_EVIDENCE → gather the missing proof and re-challenge. Only a KEEP that survives the
challenger is handed to the ideator.A scorecard built only on the user's say-so is a guess in a table. Where a parameter turns on a fact about the world, go check it.
When verifying demand (A), market size (B), willingness-to-pay / pricing (C), and
competition (D), use WebSearch / WebFetch (built-in, always available) — and the Fetch MCP
(mcp__fetch__*, shipped, when approved) — to pull real signal: search-volume and forum complaints for
the pain, competitor pricing pages and feature lists, and market-sizing references. Cite what you
find in the scorecard's probe/evidence column, and let evidence move a mark (especially toward ❌).
If web tools are unavailable, reason from the user and say so — flag the unverified marks as open
questions rather than presenting a guess as fact. Evidence-gathering never blocks the dialogue; it
sharpens it.
ideator plugin is installed, hand the opportunity to it (/ideate) for
refinement into the IDEA package. If absent, write a markdown opportunity brief to
doc/opportunities/<slug>.md and tell the user to install ideator (or run FOUNDRY's inline ideator)
to refine it. (Rich, illustrated opportunity briefs — scorecards as tables, market-sizing charts —
are produced by invoking pressroom's /publish by capability when pressroom is installed; degrade
to clean markdown when it is absent, and say so.)/discovery-goal and propose again.Carries the KAIZEN self-improvement covenant (../../knowledge/covenant.md).
When the same kind of weak candidate keeps surviving to step 5, that is not a per-scan slip but a
parameter or kill-threshold that needs sharpening — flag it for the self-improve skill so a PR
lands the fix for every future scan.
When a candidate earns a KEEP verdict (a validated OPPORTUNITY ready for refinement), and the i2p plugin is installed, mark the DISCOVER phase done so the marketplace product lifecycle and the status line advance to IDEATE:
/i2p-lifecycle done DISCOVER # order-safe & idempotent — a no-op unless a lifecycle is running at DISCOVER
Degrades silently when i2p is absent. The canonical model is i2p/knowledge/product-lifecycle.md.
npx claudepluginhub agentic-underground/idea-to-productionProvides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Searches MemPalace before answering questions about past work, people, projects, or prior decisions. Returns verbatim stored content instead of guessing from model memory.