claude

stark-constitution

Project setup, decision records, and session priming — scaffold a dev-docs/standards structure (template/backfill/upgrade modes), record architecture decisions (ADRs) under docs/adr/, write anything an agent reads and put each line where it loads (CLAUDE.md, AGENTS.md, .claude/rules, skills, memory), review agent files and memory notes clause by clause against the agent-doc playbook (.claude/rules, CLAUDE.md, AGENTS.md, skills, memory), and assign a session persona. Maps to spec-kit's constitution phase.

0 on GitHubby 21 Stark AIv0.9.0updated todaysetup
/

Plugin stark-constitution from marketplace bifrost in 21StarkCom/bifrost. Commands read from the repo's manifest — today.

What it installs

Skills (31)

  • agnes

    Run one ticket solo and unattended, with no Gru: carry it end to end through the repo's ticket → PR → review → merge → close spine, confirm the merge and the close yourself, comment the evidence on the ticket, and tear your own tab down.

  • c4-design

    Run the floorplan design loop for a C4 model: propose a LikeC4 diff, check, render, show it in the studio, accept it. Use to change, review or first model a product's architecture, C4 model or diagrams, or when a task names floorplan, LikeC4, floorplan scan, R9, a flow, a logo or a deployment view.

  • gru

    Gru drives an epic or a list of tickets to done with one Minion per ticket. Use when the operator hands over several tickets to be worked in parallel and carried through merge and closure.

  • lucius

    Open a Lucius brainstorm in a cmux tab of its own with `hermod lucius`, then stop. Lucius is a standalone app, so this skill only launches him and never runs him in this session.

  • minion

    Act as a Minion launched by Gru or the operator: own one ticket, carry it through the repo's ticket → PR → review → merge → close spine, and report the outcome to its leader: a peer such as Gru through native messaging when available, or, under `leader operator`, the operator in its own tab.

  • stark-adr

    Record and manage Architecture Decision Records (ADRs) under a repo's docs/adr/. Use for: new ADR, log/record a decision, "create an ADR", supersede an ADR, list ADRs, "why did we decide X". Wraps `brain adr` — MADR-lite, auto-numbered, repo-relative.

  • stark-agent-docs

    Use when writing or editing anything another agent will read (a CLAUDE.md, AGENTS.md or CLAUDE.local.md, a .claude/rules file, a SKILL.md or a skill's description, a doc an instruction file points to), when porting rules from Cursor or Copilot, or when deciding where an instruction, guardrail, conve

  • stark-author

    Stage 1 — write a spec and its plan in one session: tier check, short recon, plain-language interview (only what only the operator knows), one self-contained doc, one zero-context advisory pass, sign-off, commit-pinned handoff. Use to author, spec or plan a change.

  • stark-blog-sharpen

    Adversarial editor for long-form posts (blog, personal-site essays, longer internal notes) in Aryeh's voice. Use after drafting OR expanding a post, or when asked to "make this post matter / cut the padding / make it sound like a real person wrote it / does this read AI / tighten this essay". Two jo

  • stark-build

    Stage 2 — build an accepted stark-author spec: one fresh headless session per task, gated by checks the agent cannot edit, evidence and a commit per green task, a held-out e2e gate, one cross-vendor advisory review, one bounded fix round, a draft PR. Use to build or implement a spec.

  • stark-bury

    Use when retiring 21Stark code into the Náströnd graveyard — burying a subsystem of a living repo (partial burial) or putting down a whole repo (full retirement). Symptoms: "bury X", "retire the <feature>", "kill this service and keep the memory", "new corpse for nastrond", "dig a grave", "tombstone

  • stark-design-tokens

    Build, name, theme and govern design tokens: the three tiers, naming, OKLCH scales, DTCG, Style Dictionary, dark-mode and multi-brand theming, CSS variables, WCAG contrast gates and token semver. Use to make or restructure a design-token system, a tokens package or a theming layer.

  • stark-fresh-eyes

    One-shot zero-context review of a prompt, research brief, spec, or doc before it ships: a single read-only subagent re-verifies every checkable claim by a DIFFERENT method and reports defects only. Findings return to the author once — no second round, no loops. Use for fresh eyes, review prompt, che

  • stark-gha-cost

    Diagnose and cut GitHub Actions and GHAS bills for an org or enterprise: find where the money goes (product, repo, workflow, run), apply the best levers by PR, and escalate wrong billing to Support. Use for a high GitHub bill, slow or costly CI, runner choice, caching, storage or GHAS seats.

  • stark-handover

    Use when pausing or splitting work across sessions — before /clear, when context runs low, end of day, switching tasks — or when resuming after one. Triggers: "handover", "save context", "save progress", "resume", "continue where we left off", "what was I doing". Persists a numbered handover chain +

  • stark-init-docs

    Scaffold dev docs structure (template/backfill/upgrade/clean modes). Use for init docs, setup docs, scaffold docs.

  • stark-jury

    Run one post skill (stark-voice, stark-story-edit, stark-blog-sharpen or stark-story-judge) on claude, codex and gemini at once, check each result against the skill's iron rules, then merge the edits or write the calibration report. Use for a panel or best of three on a post. It never publishes.

  • stark-logging

    Use when adding, changing, or reviewing application logging — instrumenting a service/CLI/connector, choosing log levels (DEBUG→FATAL), writing structured logs, or judging whether log lines in a diff are actually useful. Symptoms: fmt.Sprintf/printf into a message, everything at INFO, log-and-return

  • stark-memory

    Audit and tidy Claude Code auto-memory so it stays under the load and recall caps. Use for memory tidy, memory audit, MEMORY.md too big, shrink the memory index, split over-cap memory files, move cross-repo facts. - claude

  • stark-persona

    Assign a famous character persona for the session with weighted random selection. Use for persona, character, voice. /stark-persona.

  • stark-refactor-plan

    Plan a codebase refactor without touching code. Inspect a repository and produce two artifacts — REFACTOR_PLAN.md and REFACTOR_BACKLOG.json — an evidence-based, phased, file-by-file restructuring plan another agent can execute. Use whenever the user wants to refactor, restructure, reorganize, modula

  • stark-release

    Cut a release: changelog review (auto-generating from git log if [Unreleased] is empty), version bump, git tag, GitHub Release. Use for release, tag, bump version.

  • stark-rich-text

    Use before you write a Slack message, a Jira description or comment, or GitHub text (a PR body, an issue, a comment), and when a task names Block Kit, ADF or mrkdwn. Write Markdown and run the fleet command that converts and sends it. Do not write code that builds Block Kit or ADF.

  • stark-rules-optimizer

    Review one repo's agent files (CLAUDE.md, AGENTS.md, rules, SKILL.md) or the auto-memory notes clause by clause against the agent-doc playbook: flag each clause GREEN, YELLOW or RED and write its redline. Read-only by default; --apply fixes them. Use to review, audit or trim agent files.

  • stark-session

    Session start (context, git state, briefing) and end (tests, merge, push). Use for session start/end, catch me up.

  • stark-story-edit

    Full-pass storytelling editor for long-form posts (blog, essays, site copy) in Aryeh's voice. Use when a draft has the facts but not the story - reads flat, buries its hook, over-explains, lands on a maxim, or needs humor and narrative pull - or when asked to "make this land / tell it as a story / p

  • stark-story-judge

    Zero-context reader verdict on a long post: cold judges, one per vendor, grade the reading on an anchored rubric with quoted evidence, plus optional audience lenses. Use before publishing or after an edit pass, for fresh eyes or a grade. It never edits the post.

  • stark-terraform-review

    Multi-agent code review of Terraform / OpenTofu (HCL) — modules, root configs, .tf/.tfvars/.tftest.hcl — for security, correctness, state safety, module-contract quality, and testing gaps. Runs the review across one or more configurable LLMs (claude/codex/gemini), each as its own subagent, then merg

  • stark-terragrunt-review

    Multi-model review of Terragrunt code (terragrunt.hcl, root.hcl, stacks, units, includes, dependency, generate and remote_state blocks) for dependency order, state isolation, mock-output safety and HCL traps. Use to review or audit a Terragrunt repo. Module HCL goes to stark-terraform-review.

  • stark-ticket

    Use when about to write, file or rewrite a ticket: before `alfred task start`, `alfred task new` or `alfred task edit --desc-file`, when asked to open, file or spawn tickets (STARK in ClickUp, or EI in Jira), a bug or a follow-up, or when turning a review finding, a plan or a request into a ticket.

  • stark-voice

    Use when drafting a Slack message, DM, reply, or short internal note as Aryeh (Kiovetsky) or "in my voice / my style / how I'd write it" - especially to Evinced / Israeli colleagues, reports, or peers. Covers Hebrew⇄English code-switching, message bursting, tone, and his feedback/praise/decision for

View source on GitHub

More from 21StarkCom/bifrost

All 11 plugins →

Zero-context fresh-eyes review of specs and briefs, multi-agent Terraform/Terragrunt (IaC) review, repository refactor planning, and application-logging authoring/review guidance. Maps to spec-kit's analyze phase.

31 skills

0

Design-token architecture — build, name, theme, and govern a design-token system the way the field's canonical sources converge on it. Covers the three-tier model (primitive/semantic/component), naming grammar, OKLCH color scales, DTCG authoring, Style Dictionary pipelines, dark-mode/multi-brand/density theming, CSS-variable distribution, WCAG contrast gates, and token semver/deprecation. Also the floorplan C4 design loop: propose a model change, check it, render it, show it in the studio, ask on the canvas with Huginn annotate, and accept it.

31 skills

0

Implement from an accepted spec. /stark-build runs one fresh headless session per task gated by checks the agent cannot edit, with a held-out e2e gate, one cross-vendor advisory review, and one bounded fix round for medium+ findings.

31 skills

0

stark-ops

claude

Repository operations beyond the spec-driven pipeline — coordinate Minions with Gru, cut releases, run session start/end, hand off work across sessions, run one ticket solo and unattended with Agnes, open a Lucius brainstorm in a tab of its own, and send rich text to Slack, Jira and GitHub with the fleet commands, with no code that builds Block Kit or ADF.

31 skills

0

Plan work before it starts — author a human-gated spec with its task DAG, and write short tickets someone else can pick up, with no guessed claims.

31 skills

0

Long-form writing craft, from the author's reusable voice profile through story editing, adversarial sharpening, independent judging, and multi-model jury calibration.

31 skills

0

Similar plugins

MetaHarness integration — surfaces score/genome/mint/mcp-scan/threat-model via skills; pairs with @metaharness/router (ADR-148/149) for cost-optimal model routing; honors ADR-150 optional-augmentation constraint

13 skills · 1 command · 1 agent · hooks

ruvnet/ruflo 81,589

context7

claudecodex

Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.

1 MCP server

upstash/context7 64,709

mempalace

claudecodex

AI memory system — mine projects and conversations into a searchable palace. 47 MCP tools, auto-save hooks, guided setup.

3 skills · 6 commands · 1 MCP server · hooks

MemPalace/mempalace 60,536

DataHub development and interaction toolkit with connector planning, PR review, catalog search, metadata enrichment, lineage tracing, data quality management, and connection setup skills

13 skills · 8 commands · 4 agents

anthropics/claude-plugins-official 47,076

ifixai

claudecodex

Independent auditing for AI agents, open source. Claude discovers your setup, confirms the cost, and audits your agent against the one question other tools skip: is it doing the job it's supposed to do, given your business rules and org structure? Test any model (Anthropic, OpenAI, Gemini, Azure, Bedrock, …) graded by self, one independent judge, or a cross-vendor panel; each provider is billed to its own account via a key in your settings, never the command line. Results render as an interactive artifact (with a static-report fallback).

1 skill · hooks

ifixai-ai/iFixAi 45,158

cicd-automation

claudecodex

CI/CD pipeline configuration, GitHub Actions/GitLab CI workflow setup, and automated deployment pipeline orchestration

4 skills · 1 command · 5 agents

wshobson/agents 42,461

Maintain stark-constitution?

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 plugins11 plugins