product-discovery is agent-read markdown (skill) from clickety-clacks/tightbeam: Drive an undefined or fuzzy product to a written spirit — elicit the real problem behind asks, reframe, deconflict, and land it in the spec's Spirit section. Use when spirit is missing, thin, or contradicted..
Indexed from public GitHub and served as immutable, content-addressed versions. Install it pinned to an exact SHA-256 with the mdr CLI, and every file is verified before it reaches your agent: the main file against the SHA-256 recorded here, the others against the git hashes of its source commit. The deterministic audit below grades the latest version, and the same checks always give the same file the same grade.
What the file says
The goal is a Spirit section the user recognizes as THEIR intent, sharper than they
said it. Conversational, not interrogation; a few questions per exchange, never a
form.
1. Start from the problem, not the solution: whatever they asked for, find what it is
FOR — "what does that get you?" / "what happens today without it?" Walk the whys
until the answer is a situation in their world, not a feature in yours.
2. Watch for the XY trap: an ask for X that serves unstated Y. When you suspect it,
name it gently — "you asked for X; it sounds like the underlying need is Y — is
that right?" The reframe-and-confirm loop is the core move: your words, their
correction, repeat until they say "yes, that's it."
3. Decompose fuzzy wants into outcomes: "how would you know this worked?" Concrete
success signs become the outcomes list; things they explicitly don't care about
become non-goals (ask for those too — non-goals prevent more waste than goals).
3b. WALK THE QUALITY AXES, lightly (ISO 25010's nine, plain words): correctness (does
the right thing), performance, compatibility (plays well with what they have), UX /
…
Pin to a label to follow the author's releases, or to a sha256 for exact bytes. Either way the resolved hash is written to mdr.lock, and mdr install fetches those bytes again and checks them, so it installs them exactly or fails.
GET https://markdownregistry.com/api/v1/artifacts/art_e5nabrfuy42xeskm
GET https://markdownregistry.com/api/v1/resolve?ref=clickety-clacks/tightbeam/product-discovery
GET https://markdownregistry.com/api/v1/blob/90fdf189a96b3d6280d3592c3796a42165d7af235207b39756d809e99a60ea49
Your agent does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
clickety-clacks/tightbeam · priv/kungfu-template/skills/<name>-example/SKILL.md · Starter skill for the <name> kungfu; replace its body and description before election.
clickety-clacks/tightbeam · priv/kungfu/agentic-engineering/skills/committing-and-pushing/SKILL.md · Commit and push discipline — build before committing, one concern per commit with behavior and structure never mixed…
clickety-clacks/tightbeam · priv/kungfu/agentic-engineering/skills/drafting-requirements/SKILL.md · The craft of a single requirement — verifiability, atomicity, active voice, banned vague words, examples as acceptance…
clickety-clacks/tightbeam · priv/kungfu/agentic-engineering/skills/feature-cycle/SKILL.md · The loop that takes a feature from intent to a user-verified result — spec, adversarial spec review, decompose…
clickety-clacks/tightbeam · priv/kungfu/agentic-engineering/skills/human-communication/SKILL.md · Write to a HUMAN in controlled plain English (derived from ASD-STE100) — active voice, simple tense, one instruction…
clickety-clacks/tightbeam · priv/kungfu/agentic-engineering/skills/model-release-intake/SKILL.md · Ceremony for taking in a newly released (or retired) model — detect catalog drift, characterize the new mind, update…
clickety-clacks/tightbeam · priv/kungfu/agentic-engineering/skills/recon-first-investigation/SKILL.md · Establish the truth before anyone fixes — reproduce, enumerate competing hypotheses, disconfirm rather than confirm…
clickety-clacks/tightbeam · priv/kungfu/agentic-engineering/skills/recon-lifecycle/SKILL.md · A recon session answers one decidable question at decision grade, in one of four forms with its confidence, records it…
clickety-clacks/tightbeam · priv/kungfu/agentic-engineering/skills/review-for-completeness/SKILL.md · Hunt what is missing, not what is wrong: the behavior, edge, error path, test, or teardown the MVP of the ask must…
cbrock84/headcount · plugins/product/skills/product-discovery/SKILL.md · Finds out whether a problem is real and a solution would work, before building it — recruiting the right people…
majiayu000/spellbook · skills/product-discovery/SKILL.md · Product discovery and market research expert. Use when validating product ideas, conducting market research, user…
magnus919/agent-skills · product-discovery/SKILL.md · Discover product requirements from human stakeholders — map who to talk to, ask questions that surface hidden…
nahid-sparktales/agent-dispatcher · skills/product/product-discovery/SKILL.md · Find the real problem behind a feature request before anyone scopes a solution — whose job is stuck, what evidence…
teambrilliant/dev-skills · skills/product-discovery/SKILL.md · Validate whether a product idea will actually work before committing engineering — find what you don't know, design the…
