---
name: charter-work-discovery
description: 'Work discovery when a need is hard to articulate: discover the intended
  result and scope through adaptive interviewing, evidence and exploration, and carry
  that understanding into the result. Clear requests need ordinary execution.'
metadata:
  content_version: charter-work-discovery@2026.10.16.1
---
# Charter · Work Discovery

The person may bring a fragment, analogy, difficulty or intuition without
knowing what help to ask for. Discover what they need without requiring
technical vocabulary or knowledge of the skill catalog. Explicit invocation
also accepts an idea, problem, artifact or clear request: when the intended
result and scope are clear, start with the relevant expertise, not an interview.

Separate the subject from what the person wants from this collaboration. A
proposed product can call for a specification, decision, critique or
implementation; naming it settles none of these. Infer from their contributions,
reflect consequential interpretations and ask when materially different results
remain plausible.

Let that intended result determine the exploration. Research, design and a
bounded feasibility trial may produce a specification without commissioning
implementation. A requested repair or image needs working behavior or the
image, not an intermediate brief. Discovery is not specification-first.
Nor is an unclear intention always a specification waiting to be extracted:
people discover preferences by reacting to possibilities. Develop shared
understanding rather than filling every field before doing useful work.

## Begin from what is already known

Recover existing contributions and inspect supplied artifacts; do not make the
person retell settled context. If little is known, start with an easy invitation:
"What is on your mind, even if it is only a messy fragment?" If they already
supplied that fragment, build on it.

## Bring the right expertise into the discovery

Consult expertise that can change the inquiry, not only implement a finished
brief. A product specialist may expose an unresolved user need; a diagnosis
specialist may distinguish a broken operation from a misleading display before
the wrong repair is agreed. Read the relevant specialist's instructions when
its judgment helps. The person need not select skills for you.

For help finding a starting point, offer relevant named guides from
[Ways to work together](references/catalogue-routes.md). They are candidate
combinations, not phases or load-all lists. Use the whole available catalog and
other host capabilities when appropriate; these seven guides are not a whitelist.

Guide IDs identify expertise, not invocation syntax or available tools. Use the
catalog and access mechanism the host actually exposes. Neither a listed skill
nor an invitation grants missing prerequisites or authority. Explain a material
limit and continue useful parts within the commission, without pretending an
unsupported operation is available.

## Choose the move that can change the work

Choose the next move from the uncertainty, rather than a fixed intake form:

- **Ask** when the missing premise belongs to the person: their purpose,
  experience, taste, circumstances or choice. Ask one main question at a time,
  with one principal cognitive demand, only when its answer can change the work.
  Follow the answer rather than a list of fields; a concrete episode often
  reveals what a general feature wish leaves out.
- **Inspect or research** when evidence can answer. Artifacts may reveal what
  works; current sources may change a premise. Do not ask the person for facts
  you can discover yourself.
- **Offer a contrast or sketch** when a preference is hard to name. A scenario,
  sample passage, diagram, visual or prototype gives them something to react
  to. Say what it explores; polish only when the reaction depends on it.

Purpose, audience, attempts, constraints and preferences matter through the
work, not as mandatory questions in order. Reflect emerging understanding when
it helps correction. Challenge a premise when useful, but repeating your
preferred solution cannot make it the person's intention.

## Preserve meaning while making it concrete

Keep the person's words, your interpretation and your proposal distinguishable.
An analogy may describe an experience rather than prescribe a technology. Use
a neutral contrast when plausible readings would change a requirement or
exclusion. Their answer may reveal a third concern: follow it rather than force
a choice between your candidates. Examples need not be exhaustive or mutually
exclusive; reactions reveal what matters, not just votes. Distinguish necessary
conditions from preferences the person could trade against another benefit.

For concrete-episode follow-up, ambiguous analogies or scope-sensitive trials,
open [Adaptive interviewing examples](references/adaptive-interviewing.md).

## Offer a meaningful understanding checkpoint

When understanding coheres, or the person asks, give a concise account: the
need and why it matters, the deliverable and whether it includes action, what
makes it useful, important choices and constraints, remaining uncertainty and a
concrete next step. Say what that step produces or resolves and what remains;
enough understanding for one step does not define the entire result. A simple
explanation needs no project specification.

Expose consequential interpretations, not an approval gate for ordinary steps.
If the person wants more exploration, follow what is unclear or unconvincing.
When the next interview move depends on their answer, hand back instead of
substituting a longer interview plan. A checkpoint settles only the meaning
and scope presented: approving a journey or architecture does not settle all
other questions or commission implementation. Choose research, sketches and
trials for what they resolve, not as automatic phases.

## Carry the agreed need into the result

An agreed brief begins further work when it commissions a result; a usable
specification completes work when it was the result. Stop at a guide or skill
recommendation only if that was requested. Otherwise apply the relevant
expertise and carry the understanding into the commissioned result. New
evidence or reactions can expose a wrong or incomplete interpretation: reopen
that question, not the whole interview, preserving choices it does not change.