Build · Create · Engineering

Build an AI Coding Session Brief and Context File

The founder gets a structured session-brief template plus a populated project-context document so each AI coding session executes against fixed architectural guardrails and logs what it changed, preventing the compounding drift the book warns about.

You receive: A JSON object: { projectContext: { patternsToFollow: [string], dependenciesToAvoid: [string], tradeoffsAccepted: [{ tradeoff, rationale }], outOfScope: [string] }, sessionBrief: { contextDocRef, taskForThisSession, constraints: [string], filesInScope: [string] }, sessionLogTemplate: { built, decisionsMade, assumptionsIntroduced } }

Part of Choose Business Model

What's verified: STUD verifies the brief and context document are structurally complete (required sections present, minimum counts met, tradeoffs each carry a rationale, log template has its three fields). STUD does NOT verify the patterns are good architecture, that the dependencies-to-avoid are the right ones, or that following the brief produces working code. It checks the scaffold's completeness, not engineering correctness.

{
 "buyerParamKeys": [
  "requireSessionLogFields"
 ],
 "checkNames": [
  "format_parses",
  "patterns_min",
  "deps_min",
  "tradeoffs_min",
  "brief_complete",
  "constraints_present",
  "no_dup_patterns",
  "log_fields"
 ],
 "constants": {},
 "deliverableLabel": "claude_code_session_brief",
 "documentKeys": [
  "constraints",
  "contextDocRef",
  "dependenciesToAvoid",
  "patternsToFollow",
  "projectContext",
  "rationale",
  "sessionBrief",
  "sessionLogTemplate",
  "taskForThisSession",
  "tradeoff",
  "tradeoffsAccepted"
 ],
 "enumLiterals": [
  [
   "built",
   "decisionsMade",
   "assumptionsIntroduced"
  ],
  [
   "yes",
   "true",
   "1"
  ]
 ],
 "submission": "a JSON document (submitted as a string), validated as data; every check below must pass",
 "tier": "doc-validator"
}