adr-critique

Guides you through seven-phase ADR drafting via focused dialogue, asking one question per turn and pressure-testing your reasoning before self-critique. Enforces specificity on components, deciders, and architectural characteristics, then routes to discovery if foundations are shallow.

ADR Discovery walks you through pre-flight context gathering before drafting an architecture decision record. It scans your repo for signals, confirms your domain and components one at a time, maps relationships, and classifies existing ADRs—outputting a human-verified discovery brief and an open-questions register. Hand off to adr-drafting once five required confirmations are locked.

c4-model produces LikeC4 specifications that enforce Simon Brown's canonical C4 model, generating context and container views alongside architecture decisions. It scaffolds the necessary DSL files and validates against a strict checklist, refusing component views, dynamic diagrams, custom element kinds, and nested systems that fall outside canonical C4 scope.

ADR Drafting helps teams document significant architectural decisions by creating new Architecture Decision Record files with a standard structure. It inspects your repository for existing ADR conventions, then drafts a new record with Title, Status, Context, Decision, and Consequences sections—defaulting to `docs/architecture/adr` when no convention exists.

adr-backfill lets you record decisions that shipped long ago but were never written up as ADRs—typically surfaced by doc audits or code archaeology. It enforces stricter honesty than forward drafting by requiring a visible backfill marker so future readers know the record was reconstructed, not contemporaneous. The skill demands two independent evidence sources and refuses to draft when confidence in the reasoning is too low.

doc-adr-fixer consumes audit reports and patches ADRs to resolve structural, referential, and content issues identified during quality review. It supports both team-mode validation (with multi-lens dispatch) and single-pass remediation, backing up originals and reverting regressions to prevent silent failures.