claudna

Plugin claudna from marketplace Claudfather in Claudfather/clauDNA. Commands read from the repo's manifest — today.

Skills (39)

• adversarial-review

  Use at any juncture where you have a plan, design, or proposal and want it stress-tested before execution — pre-mortem analysis, assumption mapping, hidden failure modes, and edge cases surfaced before they are locked in. For the standing multi-lens hardening loop on a published plan, use /claudna:i

• audit

  Use for codebase audits — security vulnerabilities, technical debt, stale documentation, visual and UX design, frontend performance (janky scroll, slow loads), cross-cutting interfaces, data model fit and redesign, scale/throughput and multi-tenant isolation, birds-eye repositories, whole-system com

• build-all

  Use when executing multiple phase or plan docs as one ordered workstream — an epic spanning several PRs, possibly several repos, with phases that must land in a declared order. Also when a forge epic's phases are ready for execution. For a single plan doc, use /claudna:build instead.

• capture

  Use to save knowledge to the vault — a note you write, external content (an article URL, a file, a transcript), or the current session's learnings (bare `/claudna:capture` distills what happened after corrections, surprises, or fixes, before compaction). One write door. To read the vault use /claudn

• claudron

  Use to search the shared fleet vault by term, check its health, or diagnose and migrate it — the lookup, status and doctor verbs over the Claudron CLI. Requires the Claudron CLI. To save knowledge to the vault (a note, external content, or the current session's learnings), use /claudna:capture; to r

• cleanup-legacy-install

  Use when upgrading from a pre-plugin clauDNA install (install.sh, claudefather, or claudfather) to the marketplace plugin. Detects and removes the legacy ~/.claude/ overlay so plugin-installed skills don't get shadowed by stale copies.

• dbt

  Use when you need to run dbt commands against Snowflake -- build, test, compile, or full-refresh.

• delegate

  Use when delegating a bounded task between Claude Code and native Codex. For same-provider work or isolation, use /claudna:worktree.

• development-retro

  Use when a PR has been merged, a development session is ending, or the user asks what else you noticed — surfaces journey-specific observations, systemic patterns, and friction points from the implementation process.

• extract-system-spec

  Use when you need a portable, anonymized reconstruction spec for a whole system (or a single wing or targeted slice) — a standalone rebuild blueprint an independent team could implement in a fresh session without the original repo. Exhaustive and expensive by design: reserve it for when a full rebui

• file-github-issue

  Use when a user wants to file a GitHub issue with a screenshot against a GitHub repo. Expected inputs: a screenshot, a short description, the target repo, and the page URL or app section where the issue was observed (the skill prompts for the URL if missing).

• find-skills

  Use when the user asks about finding, discovering, or installing agent skills from the public ecosystem, or asks how to do something that might exist as an installable skill. For raiding a specific repo's patterns into clauDNA, use /claudna:heist.

• forge

  Use when planning a workstream that spans multiple PRs, involves decision forks, or needs structured phasing before implementation — a technical spec, design document, initiative roadmap, or pre-implementation proposal. Not for single-PR changes with an obvious path; plain plan mode covers those.

• github-activity-report

  Use when the user asks for GitHub activity stats (PRs authored, commits, top contributors, top repos) across an organization or their personal account over a time window. Iterates per-repo to avoid the search API's 1,000-result cap.

• heist

  Use when you want to raid a GitHub repo for skills, config patterns, or novel approaches worth adopting into clauDNA. For browsing or installing published skills from the open ecosystem, use /claudna:find-skills.

• index

  Use after creating, promoting, or editing docs in a directory whose INDEX.md should be regenerated, when frontmatter needs validating against the documentation standard, or to audit knowledge-base health. The sole writer of INDEX.md files.

• init-project

  Use when setting up a new project or adding shared project instructions (CLAUDE.md or AGENTS.md), CHANGELOG.md and documentation to an existing project.

• investigate-app

  Use when something is broken in production and you need to investigate -- errors, outages, degraded performance, or unexpected behavior.

• ironclad

  Use to harden a plan or review a PR with a panel of independent lenses, or to step back and challenge a plan's premise (are we solving the right problem?) — primary target a §4.1 plan Issue produced by /claudna:forge; also reviews implementation PRs. For a live stress test with no plan to review, us

• kindle

  Use before any creative work — creating features, building components, adding functionality, or modifying behavior. The ideation phase: discovers intent and requirements into an approved spec before /claudna:forge plans and /claudna:ironclad hardens. Requires a presented design and user approval bef

• modal

  Use for Modal serverless operations — deploy or roll back an app, stream, search, or debug logs, or check workspace status (apps, containers, secrets, volumes). Replaces /modal-deploy, /modal-logs, /modal-status.

• neon

  Use for Neon PostgreSQL operations — query (run SQL, explore schema), branch (create, list, delete, or reset database branches for safe experimentation), or info (connection and size overview). Replaces /neon-branch, /neon-info, /neon-query.

• product-enhance

  Use when you have specific product issues to triage or want to discover enhancement opportunities in an existing product through gap analysis. For exploring what a codebase could become, use /claudna:product-vision.

• product-vision

  Use when you want to explore what a codebase could become — candidate features one or two hops from existing infrastructure, compound plays, and a trajectory aligned to the project mission. For triaging known issues in an existing product, use /claudna:product-enhance. For discovery gated on an acti

• promotion-intake

  Use when a Claudosseum promotion package (manifest + battle history + telemetry + SKILL.md bundle) is ready to land in clauDNA — validates it against the promotion contract and stages the skill as a reviewable PR. Not for authoring new skills by hand; use /claudna:skill-scaffold for those.

• publish

  Use when a finished markdown document — plan, audit or review findings, retro, decision, or knowledge page — needs to reach its destination: the shared-docs vault, the repo's documentation/ tree, a GitHub issue, a PR description, the chat session, a Notion page, or the default private session artifa

• qa

  Use for headless-browser QA against a running app — systematic bug-hunting with optional fixes, post-deploy canary monitoring, performance benchmarking, ad-hoc browser interaction, cookie import, or an autonomous full-site visual crawl with per-route findings and explicit issue export. For a judgmen

• railway

  Use for Railway operations — deploy or roll back a service, view, filter, or debug logs, or check project status (services, deployments, environments, metrics). Replaces /railway-deploy, /railway-logs, /railway-status.

• recall

  Use before starting substantive work in a repo — surfaces what the fleet and this project already know (prior decisions, patterns, conventions) so you don't rework solved problems. Orientation briefing over `claudron recall`; falls back to scanning INDEX.md when Claudron is absent. First verb in the

• review-work

  Use when work needs review before it lands — uncommitted changes before committing, a pull request before merging, or several PRs reviewed in parallel. For hardening a plan Issue with the lens panel, use /claudna:ironclad. Replaces /review-changes, /review-pr.

• session

  Use at session boundaries — resume at the start of a new session (resolve a scoped handoff and brief on where to pick up), handoff at the end of a session (write it), checkpoint for a mid-session save without the full ceremony, or name to label the session. Also use to look back at recorded sessions

• ship

  Use when changes are ready to land — a fast local checkpoint, or the full flow through push and an open pull request. Defaults to the PR flow; say 'commit' or 'quick commit' for a checkpoint-only. Replaces /commit-push-pr, /quick-commit.

• skill-scaffold

  Use when you want to scaffold a new clauDNA skill directory with correct structure, valid frontmatter, and a starter body that passes validation.

• sync-branches

  Use to bring every local branch and worktree across your sibling git repos up to date with origin/main in one pass — fetches, fast-forwards where safe, and reports conflicts without forcing anything. For creating a NEW worktree to isolate feature work, use /claudna:worktree instead.

• using-claudna

  Use to verify clauDNA installation is healthy — hooks, skills and capture — or orient at conversation start and choose an installed skill for a task. For discovering third-party skills from the public ecosystem, use /claudna:find-skills. Replaces /skill-health.

• vercel

  Use for Vercel operations — deploy to production or preview (or roll back), view, filter, or debug deployment logs, or check project status (deployments, domains, env vars, config). Replaces /vercel-deploy, /vercel-logs, /vercel-status.

• verify-completion

  Use when about to claim work is complete, fixed, or passing -- requires running verification commands and confirming output before making any success claims.

• weigh-development-paths

  Use when at a development junction with multiple viable approaches — architecture choices, refactoring strategies, where to put new code, or which pattern to follow. Triggers on 'which approach', 'how should I', 'Option A vs B', or any point where the next step isn't obvious and the wrong choice cre

• worktree

  Use when starting feature work that needs isolation from the current workspace, or when running multiple Claude sessions in parallel on different branches — creates and manages git worktrees. For fetching latest main into every existing worktree and branch across your sibling repos, use /claudna:syn

Subagents (8)

• code-reviewer

  Code quality reviewer. Evaluates implementation for correctness, clean design, test coverage, and maintainability.

• dbt-engineer

  Analytics engineering agent. Writes, reviews, and tests dbt models against Snowflake.

• modal-ops

  SRE agent for Modal infrastructure. Diagnoses production issues with serverless GPU workloads.

• neon-analyst

  Data analysis agent for Neon PostgreSQL. Queries via psql and provides insights.

• railway-ops

  SRE agent for Railway infrastructure. Diagnoses production issues and analyzes deployments.

• snowflake-analyst

  Data analysis agent for Snowflake. Queries via snowsql and provides insights.

• spec-reviewer

  Spec compliance reviewer. Verifies implementation matches plan requirements exactly -- nothing missing, nothing extra.

• vercel-ops

  SRE agent for Vercel infrastructure. Diagnoses production issues and analyzes deployments.

Similar plugins