claudecodex

cogniva-dev

Cogniva's development toolkit: pre-planning idea exploration (explore-idea), technical experiment design, recording and review (investigate), human-confirmed architectural decision records with provenance and relitigation weight (adr), deferred work capture and grooming (backlog, groom-backlog), low-involvement work selection (easy-work-scan), repo scaffolding (repo-init, add-module), the feature lifecycle (plan-feature, execute-feature, quick-fix) with optional per-clone isolated-worktree execution and ledger-based cleanup (cleanup-work, cleanup-allwork), Module dependency-graph generation (module-deps), and read-only status roll-ups (feature-status, module-status, repo-status, workflow-status).

1 on GitHubby Cognivav0.11.0MITupdated 10 days ago
/

Plugin cogniva-dev from marketplace cogniva in cogniva/cogniva-skills. Commands read from the repo's manifest — 9 days ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.

What it installs

Skills (21)

  • add-module

    Use when adding a new Module to a repo on the Module bundle layout (src/Modules/<Name>/ with Contracts/Domain/Application/Infrastructure/UI projects, optional Client) - scaffolds the projects, wires references and tests, updates the glossary. Not for repos that follow the dotnet profile's default ki

  • adr

    The authority on Architecture Decision Records (ADRs) — their format, provenance, relitigation weight, and numbering. Use to record a confirmed architectural decision as an ADR (ALWAYS with explicit human confirmation first), or to check how open an existing decision is to being reopened. Also runs

  • applicable-rules

    Discover the repository instruction chain, placement constraints, and relevant obligations for intended target paths before implementation. Read-only; never changes repository or lifecycle state.

  • backlog

    Use to capture a planned deferral - work with a stated reason not to do it now ("blocked on X", "after feature Y", a decision pending, an explicit human "later") so it is not lost. Reason-less real work is done now or routed to quick-fix / plan-feature, never parked here. Directly invokable AND the

  • cleanup-allwork

    Checkout-wide safety net that finishes EVERY cleanupable worktree on this checkout, including ones from sessions you closed or forgot. Works from the JSON ledger recipe alone - no original-session context needed. For each cleanupable record it retries integrate if queued, runs the close-out recipe (

  • cleanup-work

    Close out THIS session's worktrees after you have validated the work. Runs the close-out recipe for each worktree this session created and marked cleanupable (flips state.md Status, removes the merged+clean worktree, prunes the ledger record). Session-scoped and safe - never touches in-progress or o

  • easy-work-scan

    Use when the user wants a shortlist of backlog work that can be tackled with little to no involvement from them ("scan the backlog for easy stuff", "what can you just get on with?") - a read-only scan qualifies every open item against a five-part low-involvement test and presents the shortlist, plus

  • execute-feature

    Use to execute a feature plan from a single prompt — given EITHER a plan name/path OR the full text of a plan pasted straight into the prompt (e.g. copied from another agent). Accepts any plan format — freeform plans are converted to the task format first, so checkbox tracking and resume work for ev

  • explore-idea

    Use when the user wants to BRAINSTORM and develop an idea before any planning - a long, question-driven design conversation that produces only disposable thinking docs (gitignored .explore/), never a plan and never code. Invoke explicitly only; the goal is to deepen the idea, not to converge on acti

  • feature-check

    Provide workflow-neutral preflight, review, and commit-readiness evidence for a bounded implementation. Delegates authority discovery and mechanical gates; it never owns lifecycle mutations.

  • feature-status

    Use to see what feature work is remaining and in-flight. Read-only scan of docs/plans/<Module>/*/ plan checkboxes and state.md files plus git worktrees - reports per-feature task progress, the next actionable task, pending manual-validation (⛔) gates, active worktrees, and any integrations queued wa

  • gate-check

    Run the repository's canonical mechanical completion gates without taking lifecycle ownership. Read-only except for commands explicitly configured by the target repository's green gate.

  • groom-backlog

    Use to groom the backlog - an evidence-gathering scan proposes closures (already-done, obsolete, superseded, duplicate) for loose BACKLOG.md items and deferred stubs, and the user confirms everything once before any edit happens. Deep pass (still-wanted?, merge/split, tier/size review) is opt-in; ot

  • investigate

    Design, record, or review technical experiments within a continuing investigation, with proportionate rigor and evidence-based challenge. Invoke manually only; experiment execution needs user approval, and accepted documentation normally receives a scoped local commit.

  • module-deps

    Module bundle layout tool - regenerate the Module dependency graph (docs/architecture/module-dependencies.html + .md) from the .csproj ProjectReference graph of a repo laid out as src/Modules/<Name>/, or check it for cross-Module cycles with -Check. Use when the user asks for the module dependency g

  • module-status

    Use to see the whole state of ONE Module - in-flight features, deferred backlog stubs, and loose BACKLOG.md items in a single read-only view. Wraps the feature-status plan/state scan (scoped to the Module) and adds the Status: lifecycle, deferred stubs, and the loose-item count. No subagents, no edi

  • plan-feature

    Use when designing ONE feature with a strong model before implementation - runs a focused design session and emits a task-segmented feature plan (executable by /execute-feature). Pairs with adr (ADRs) and glossary. Starts only on an explicit user request or an approved dispatch from easy-work-scan/g

  • quick-fix

    Use for small follow-up changes (UI tweak, bug fix, copy change) without a formal feature plan. Runs the change via a background Workflow so it can be fired repeatedly without bloating context; in worktree mode the change is isolated in a git worktree and auto-integrated. Starts only on an explicit

  • repo-init

    Use when starting a brand-new .NET repo - scaffolds git, the solution, Directory.Build.props, the chosen hosts, AGENTS.md and a seed glossary, and adopts and declares the dotnet architecture profile

  • repo-status

    Use to see the whole repo at a glance - the live roadmap. Read-only cross-Module roll-up of docs/plans/ - one row per Module (features done/total, in-flight, deferred stubs, loose-backlog count), plus repo-level BACKLOG.md items and global action items (queued integrations, pending ⛔ gates, active w

  • workflow-status

    Read-only status of background Workflow runs (execute-feature, code-review ultra, custom workflows) for this repo - including ones started in OTHER sessions. Scans the on-disk workflow journals so it works from the IDE/chat with no terminal `/workflows` UI. Reports each run's name, originating sessi

Hooks (4)

  • PreToolUse

    node "${CLAUDE_PLUGIN_ROOT}/scripts/guard-primary-edit.js"

  • PreToolUse

    node "${CLAUDE_PLUGIN_ROOT}/scripts/guard-primary-git.js"

  • PostToolUse

    node "${CLAUDE_PLUGIN_ROOT}/scripts/nudge-backlog-commit.js"

  • PostToolUse

    node "${CLAUDE_PLUGIN_ROOT}/scripts/guard-module-cycles.js"

View source on GitHub

More from cogniva/cogniva-skills

All 2 plugins →

cogniva-skills

claudecodex

Cogniva's general-purpose toolkit: glossary, reference, project-requirement, and project-context skills for any project

5 skills

1

Similar plugins

Matt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).

39 skills

mattpocock/skills 323,108

superpowers

claudecodex

Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques

15 skills · hooks

obra/superpowers 316,836

nextjs

claude

Official Next.js skills: adopt and optimize Cache Components, adopt Partial Prefetching, and verify runtime behavior against a running dev server.

6 skills

vercel/next.js 143,495

Taste Skill

claudecodex

tasteskill 2.5.1: professional company websites at the figma.com and x.ai level, with one section per screen, real photographs, real product UI, effects kept for tech and gaming pages, and a real-browser checker and finish gate.

1 skill

Leonxlnx/taste-skill 114,391

AI-powered codebase understanding — analyze, visualize, and explain any project

9 skills · 10 agents · hooks

Egonex-AI/Understand-Anything 92,410

diagram-design

claudecodex

Editorial diagrams in your brand, as self-contained HTML and SVG. Architecture, flowcharts, sequence, state machines, ER and database schemas, timelines, swimlanes, org charts, Gantt, Wardley maps, user journeys, service blueprints, UML class, lifecycle phase maps, and charts from bar, line and scatter to Sankey, treemap and heatmap, plus many more. Redraws draw.io, Mermaid, Excalidraw and PlantUML files. Exports SVG and PNG.

1 skill · 7 commands

cathrynlavery/diagram-design 66,757

Maintain cogniva-dev?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins2 plugins