requirement-analyzer

Requirement analyzer

Execution kind: agent-authored. Read safety and artifact contracts. Produce an explicit, provenance-marked requirement-analysis draft; do not invent authority for unresolved requirements.

Resolve ./node_modules/.bin/qa-skill before qa-skill on PATH; if absent, stop with setup guidance. Never use remote npx. Ingest the finished draft through the local runtime.

Example — standalone analysis:

QA_SKILL="${PWD}/node_modules/.bin/qa-skill"; [ -x "$QA_SKILL" ] || QA_SKILL="$(command -v qa-skill)" || exit 1
RUN_ID="$("$QA_SKILL" run create --root . --mode plan --environment-file environment.json | node -p 'JSON.parse(require("fs").readFileSync(0, "utf8")).runId')"
"$QA_SKILL" artifact ingest --root . --run-id "$RUN_ID" --type requirement-analysis --file drafts/requirement-analysis.json

PowerShell:

$QaSkill = Join-Path $PWD "node_modules/.bin/qa-skill.cmd"
if (-not (Test-Path $QaSkill)) { $QaSkill = (Get-Command qa-skill -ErrorAction Stop).Source }
$Run = (& $QaSkill run create --root . --mode plan --environment-file environment.json | ConvertFrom-Json)
& $QaSkill artifact ingest --root . --run-id $Run.runId --type requirement-analysis --file drafts/requirement-analysis.json

run create returns the new runId and closes its process lock without finalizing the workspace, so later standalone specialist skills can keep registering canonical artifacts in that run.

Example — full run: hand the registered analysis to testcase-designer, scaffold full-workflow.json from the returned source run ID, then run it.