The skill
discussionas its source filebootstrap/skills/discussion.mdwrites it, rendered bygen-processes-viz.tsbecause it has no instruction page on this site — do not edit here. All processes
discussion — settling what an agent cannot read off disk
Skill discussion: Settle which harness, and which repositories, when no file says: the output is a document conforming to discussion.output.schema.json.
Named by: Complete initialization, Determine the harness and repositories, Initialize a harness
An agent that has just read bootstrap’s README.md knows the shape of a
harness and nothing about this one. Two questions have no answer in any
file it can reach:
- which harness this repository should become:
bootstrapitself, or any Harness built on it; - which repositories are involved — where the dependency is read from, where the new declaration is written to, and whether they are the same.
Only whoever asked for the harness can answer them, so the agent asks.
This skill says what the asking must produce.
processes/discussion.bpmn narrows the
candidates and then calls
human-agent-discussion, the one reusable way
bootstrap asks a person anything, with no default allowed.
Inputs and outputs
| in | discussion.input.schema.json: what the agent already knows, and which of the two unknowns is open |
| out | discussion.output.schema.json: harness, repositories[], outcome, determinedBy, answeredBy, and the exchange that produced it |
Both are plain JSON Schema in schemas/, readable with nothing installed.
The task is finished when a document conforming to the output schema exists,
not when a conversation has taken place.
The participant may be a person or a sibling agent that already holds the
answer; the output records which (kind: person | agent), as
human-agent-discussion
says.
First, say where you are
Before narrowing anything (A_DetermineWhere), determine the three things
human-agent-discussion
requires before any question:
- the Process: Initialize a harness, inside Determine the harness and repositories. A Bootstrapping Agent can be nowhere else (FR-2);
- the Roles: itself as the Bootstrapping Agent; the person it will ask as the Requestor. A request relayed by somebody else leaves the Requestor undetermined;
- the context: what the request said, and what the checkout holds.
One that cannot be determined is never filled with a default: it is asked about first, and the harness question waits.
Rules
Ask the fewest questions that settle it. Narrow the candidate list from
context first (a repository whose root already holds a declaration,
<name>.json, is not a blank slate) and put the rest as one question with the
candidates named. A question a reader must go and research is not ready.
Never infer the answer from a default. If nothing is known, the output’s
determinedBy is asked, not assumed. assumed is permitted only for a
default the README states, and assumption then names it.
Record repositories as pairs, not as a list. read from and written to
can be the same repository or two; the point of determining them is to know
where the declaration lands.
Never infer written-to from push access. A remote repository is
read-from until the Requestor authorizes writing to it
(confirm-harness).
A refusal is an outcome. If the participant declines, that is
outcome: "unsettled" with what was asked and what is still open — not a
guess, and not an error.