Deep Interview Skill
First step of the ChainlessChain canonical coding workflow:
$deep-interview → $ralplan → $ralph / $team
Use this skill when the user's request is vague, the scope is unclear, or the boundaries and non-goals have not been agreed upon.
It writes a structured intent.md to
<project>/.chainlesschain/sessions/<session-id>/intent.md which the
next step ($ralplan) will read.
Usage
$deep-interview "add authentication to the API"
Inputs (params)
goal(string, required) — the raw user requestclarifications(string[], optional) — follow-up answers already gatherednonGoals(string[], optional) — explicit out-of-scope itemssessionId(string, optional) — reuse an existing workflow session
Output
intent.md file path + the rendered markdown, plus LLM guidance asking
the agent to ask the user any remaining clarification questions before
advancing to $ralplan.