Normalize all starting evidence into a subject spec run. Use this when beginning a Forge spec from code, docs, transcripts, screenshots, URLs, or a sparse user request. Creates the subject slug, frontmatter, evidence ledgers, request archetype, evidence-density classification, and canonical readiness skeleton.
Use this skill first.
Turn raw inputs into a normalized evidence run with:
specs/README.md exists.specs/README.md is missing, create it with:
<!-- forge:spec-index:start --><!-- forge:spec-index:end -->specs/README.md exists but the managed block is missing, add the
managed block without deleting human prose outside it.specs/<subject>.md.specs/_artifacts/<subject>/.repodoctranscriptuiuser_statementanswerobservationevidence_unit record with:
idsource_typesource_refdirectnessauthorityfreshnessindependence_groupextraction_methodsource_origin_keys from evidence provenance, not raw file count.request_archetype:featureanalogy_featureparity_cloneintegrationbugfixmigrationrefactorreverse_specstarting_evidence_density:sparsemixeddensecritical_decision_coverage with these buckets:core_outcomescope_boundaryimplementation_constraintsdependencies_and_integrationsacceptance_signalOverviewIntake SummaryEvidence Modelmanifest.jsonrun-state.jsonbranch-registry.jsoninput-log.mdevidence-ledger.jsonlclaim-ledger.jsonlUBIQUITOUS-LANGUAGE.md does not exist at the project root, create it
with an initial set of domain terms extracted from the evidence sources.
If it already exists, read it and use existing terms for consistency.
Append any new domain terms, entity names, or concepts discovered during
intake.AGENTS.md at the project root (if it exists). Note project
conventions, patterns, and anti-patterns it encodes — these are prior
implementation learnings that should inform the new spec rather than be
rediscovered. Record relevant conventions as observation evidence units
with source_type: prior_learnings.Intake Summary stays short.specs/README.md is navigational.sparse analogy or feature runs should default to clarification if the
critical decision buckets are not explicit in evidence.