Claude Code plugins for the full feature lifecycle: idea → Value Increment → ARD → spec → design → Epics → implementation → docs → release notes, each gated by an Opus review. Tracks what every feature costs in $ by phase, role and model, and captures friction in-flight so the workflow keeps improving.
Agent
Part of dev-workflows
Reviews OpenAPI specification files against REST API and IAM permission-naming guidelines. Runs a deterministic Spectral lint against the bundled ruleset first (silently skipped when no Spectral CLI is available), then reviews what a linter cannot express. Checks version consistency, required elements, naming…
Agent
Part of dev-workflows
Reviews an Architecture Requirements/Decision Document (ARD) authored by /create-ard for grounding integrity (every as-is claim cites a real file:line), AD#N well-formedness (Binds/Prevents/testable Rule), non-contradiction of inherited PRD-level invariants, altitude purity (no per-repo solutions at PRD level), and…
Agent
Part of dev-workflows
Adversarially reviews a BRD package before it goes to the customer — attacks the position rather than summarising it, and returns [SR#n] findings each requiring a disposition. Read-only. Uses Claude Opus.
Agent
Part of dev-workflows
Extracts a requirement inventory from a customer-supplied BRD — one [BR#n] row per requirement, with a source anchor and unconfirmed defect candidates. Splits a requirement carrying more than one obligation. Read-only; never writes the source. Uses Claude Sonnet. Extraction is mechanical; defect classification is the…
Agent
Part of dev-workflows
Grounds specific BRD claims against a single code repository at a pinned commit — one [CG#n] finding per claim, with file:line evidence and a verdict from the closed set. Answers "is this claim true of this commit?", not "what capability exists?" — that is code-scanner. Read-only; one instance per repo, up to 4…
Agent
Part of dev-workflows
Post-implementation code review for SIGNIFICANT / HIGH-RISK tasks. Checks correctness, security, architecture, edge cases, migration, dependencies, tests, rollback. Returns PASS / PASS WITH RECOMMENDATIONS / BLOCK and gates the test run. Uses Claude Opus.
Agent
Part of dev-workflows
Scans a single code repository for existing capabilities and gaps relative to a set of themes. Themes may come from a Product Requirements Document / Epic (Epic writing), from an implementation spec (/implement multi-source scanning, broad-then-narrow per §8.5), from a Jira item being specified (/specify light…
Agent
Part of dev-workflows
Reads a returned customer review of a BRD package in two modes — schema mode parses a file written against the customer review schema; free-text mode drafts that same schema from prose and emits every inferred decision as an unconfirmed candidate. Never promotes an inference to a customer decision. Read-only. Model…
Agent
Part of dev-workflows
Reconciles a BRD against an exported design frame set — one [DG#n] finding per divergence, in four classes: a frame shows a field the BRD never requires; the BRD requires a field no frame shows; a frame contradicts BRD text; a frame implies a capture the code cannot perform. Read-only. Model tier assigned by the…
Agent
Part of dev-workflows
Reviews an engineering design.md authored by /design against the design-format authority and traceability to its specification.md — architecture/interface/seam/test-strategy soundness, coverage of every in-scope requirement, and decision-completeness. Treats any unresolved design.md open question as a BLOCKER.…
Agent
Part of dev-workflows
Reads a single code repository's PR diff(s) and returns a documentation-focused summary. Host-aware resolver — uses the gh CLI for GitHub when available, falls back to pure-local-git strategies for Bitbucket Cloud, Bitbucket Server, and GitHub when gh is absent. Designed for parallel invocation (one instance per repo…
Agent
Part of dev-workflows
Applies targeted fixes for BLOCKER and MAJOR findings from a doc-reviewer or epic-reviewer report, or for violations from docs-style-checker / prose-style-checker. Mirrors review-fixer for the docs domain. Returns a structured fix report; caller re-runs the reviewer. Shared between /document and /epics. Model tier…
Agent
Part of dev-workflows
Finds the right place(s) in a docs repository to write new or extended documentation for a feature. Returns a prioritised list of write targets (extend-existing, new-page-in-existing-section, new-section) with rationale. Heuristic + grep work; no content written. Model tier assigned by the caller per the model-routing…
Agent
Part of dev-workflows
Synthesises Jira data, per-repo diff summaries, and confirmed write targets into a documentation checklist the writer follows and the reviewer checks against. Detects the repo's image policy (local vs CDN-upload) and annotates per-screenshot placement. Does NOT write content. Model tier assigned by the caller per the…
Agent
Part of dev-workflows
Reviews product documentation written by /document for correctness, completeness, and fitness for purpose. Returns PASS / PASS WITH RECOMMENDATIONS / BLOCK. Uses Claude Opus. Epic drafts are reviewed by epic-reviewer (a separate agent); this reviewer is product-docs-only.
Agent
Part of dev-workflows
Writes product documentation for /document from a structured handoff file — applies the doc-planner checklist, discrepancy decisions, snippets, screenshots, frontmatter, and internal links. Write-only (no git). Returns the list of files written. The orchestrator pins it to the §2 Opus reasoning chain.
Agent
Part of dev-workflows
Read-only documentation grounding for authoring commands. Given a docs root ($DOCSPATH), a feature summary, and optional Jira key/themes, retrieves the most relevant existing product-doc pages and returns a bounded digest — docsreferences (positive grounding — same-feature facts, analogous precedents to model after…
Agent
Part of dev-workflows
Runs the docs repo's project-configured prose linter (e.g. Vale) on files written by /document (Jira mode, or direct mode) AND, when the prose-style plugin is installed, also runs prose-style-checker as a complementary semantic / cross-page-consistency pass. Merges and dedupes both finding sets into the doc-reviewer /…
Agent
Part of dev-workflows
Reviews Epic drafts written by /epics for goal clarity, acceptance-criteria testability, scope boundaries, and non-duplication with existing Epics under the parent PRD. Returns PASS / PASS WITH RECOMMENDATIONS / BLOCK. Uses Claude Opus. Product documentation is reviewed by doc-reviewer (a separate agent); this…
Agent
Part of dev-workflows
Writes child Epic-definition files for /epics from a structured handoff file — one file per Epic, following the Epic template, traceable to the jira-reader handoff and code-scanner evidence. Write-only — writes vault content, never commits (still true — it runs no git at all). Returns the list of Epic files written.…
Agent
Part of dev-workflows
Independently re-derives a grounding finding from its own source — the pinned repository for a [CG#n] or a class-4 [DG#n], the exported frame set for a design-only [DG#n] — and returns agree / extend / contradict / unprovable with its own evidence. It does NOT check citations — checking a citation only proves the…
Agent
Part of dev-workflows
Reviews app code and UI for compliance with public UI design-system and accessibility standards. Checks app header, data table, filter field, connections, permissions, settings, dashboards, accessibility/WCAG, terminology, and data naming. Triggers on 'review for guidelines', 'check compliance', 'UI guideline review'…
Agent
Part of dev-workflows
Ingests one idea source (inline prompt, a markdown file with wikilinks/images, a community post, or an exported Jira ticket — either product feedback (an RFE) or an existing Product Requirements Document the idea extends, parallels, or rewrites) from the user's vault and returns a structured source digest for /idea.…
Agent
Part of dev-workflows
Post-session maintenance agent. Reads what happened during an implementation, fix, or upgrade session and produces a structured Lessons Learned report with actionable suggestions for improving project tooling — CLAUDE.md rules, reference docs, hooks, command workflows, and new skill patterns. Suggest-only; does NOT…
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: