sunstack-dev
claudecodexShared engineering practices, principles, language standards, research, debugging, docs, operations and presentation.
43 skills
Frozen legacy v1 workflow: roles, orchestration, design, planning, implementation, review and handoffs. Requires sunstack-dev. Do not edit without explicit owner instruction.
Plugin sunstack-workflow-v1 from marketplace sunstack in ShravanSunder/sunstack. Commands read from the repo's manifest — yesterday.
implement-planUse when executing or continuing one ready plan whose delivery context authorizes implementation, or correcting the implementation-owned findings of one review. Not for plan or design defects, independent review or PR work.
implementation-handoffUse when packaging implementation state for another agent, reviewer, CLI, or future session: a copy-paste blurb, reviewer prompt, Claude/Gemini/Codex handoff, unfinished-work handoff, or post-implementation review packet.
implementation-pr-wrapupUse when pushing, opening, updating, monitoring or finishing a GitHub pull request after implementation: its description, checks, comments, review threads, mergeability or \"merge when ready\". Not for fresh code review.
implementation-reviewUse when independently reviewing implemented code, proof, a diff, commit, PR head or one bounded remediation before PR readiness, even with missing or conflicting governing inputs. Not for design review (spec-program-review).
orchestrator-designUse when asked to run, resume or finish one design cycle. Not for one direct design phase, changing a plan during delivery, implementation delivery or PR work.
orchestrator-implementation-goalUse when starting, resuming, auditing or completing an implementation delivery goal. Not for a design-only cycle or one direct phase.
plan-handoffUse when packaging an existing implementation plan for another agent, CLI, machine, or future session, especially when the user asks for a copy-paste prompt, next-agent context, or portable plan packet.
plan-implementationUse when writing the first implementation plan from a reviewed four-document design, an admitted improvement finding or a direct planning request. Not for changing it later (plan-update) or repo audits (plan-improve-repo).
plan-improve-repoUse when auditing a repository for improvement opportunities, backlog-worthy refactors or quality gaps, and turning admitted findings into plans. Not for planning from a reviewed design.
plan-updateUse when a plan's meaning changes in delivery: design correction, plan defect, slice split or merge, new dependency or proof target, PR re-split. Not for first plans (plan-implementation) or in-slice choices (implement-plan).
practices-manage-agentsUse when delegating to subagents, teams or parallel agents (fan-out too): picking role, form, model; spawning, commissioning, continuing, steering, awaiting, monitoring, verifying or assessing them. Not for doing the phase.
program-designUse when defining or revising structural How: entities' owners, homes and schemas; components, interfaces, state, flows, failure, trust boundaries, proof seams. Not for what entities mean or other Why/What (spec-design).
role-advisorUse when commissioned or named as the Advisor: a second look at a design, plan or acceptance argument, the owner's questions, or the Lead's stalled work. Not for authorship, acceptance or independent review findings.
role-implementerUse when commissioned as the Implementer for approved changes, product tests, associated proof and corrections. Not for defining the governing outcome, observing as the paired Verifier or independent review.
role-leadUse when you are the owner's main engineering partner responsible for governing design, plans, integration and acceptance. Not for a helper assigned bounded evidence, implementation, observation or independent review.
role-operatorUse when assigned as the Operator for prescribed execution, watch, scraping, extraction or guided collection under supplied criteria. Not for choosing engineering probes, synthesizing conclusions or repairing failures.
role-researcherUse when assigned as the Researcher responsible for a checked answer to a bounded question or continuing investigation. Not for prescribed extraction alone, governing design authorship or delivery acceptance.
role-reviewerUse when commissioned as an independent Reviewer for a declared artifact or its corrections, with no authoring history. Not for authoring, implementation or substituting advice for independent findings.
role-verifierUse when commissioned as the paired Verifier to build or keep the verification system, challenge proof feasibility before a build, or observe the adopted outcome. Not for product repairs, writing product tests or acceptance.
spec-designUse when authoring or revising Requirements from settled meaning, or a Specification's obligations, entities and coverage views. Not for eliciting unwritten owner meaning (discuss-pathfinding).
spec-handoffUse when packaging spec, design, architecture, or product-decision context for another agent, CLI, machine, or future session before an implementation plan exists.
spec-program-reviewUse when independently reviewing a Specification, Program Design or the four-document set, including reader understanding or readability, or whether a spec- or program-only change needs review. Not for implementation review.
verification-designUse when planning how a design's requirements will be proven, after Program Design or when that plan is missing or stale. Not for checking a delivered claim (practices-verification), building journeys or regression tests.
Shared engineering practices, principles, language standards, research, debugging, docs, operations and presentation.
43 skills
Skill authoring for any repository: design skill changes, write their files, turn their standards into checked rules, review them with independent agents, and audit which skills to create, update, merge, or skip.
11 skills
Editable v2 workflow: four design skills, four roles, and agent management. Requires sunstack-dev. Does not replace frozen v1.
9 skills
Project scaffolding with standard dev configs (biome, ruff, vitest, pytest, cursor rules, claude hooks)
1 skill · 2 commands
Agent collaboration, shared message boards, wake-ups and scheduled workflows through Codex Router.
1 skill
Use Agent Studio worktrees, pane messages, context and App commands from coding agents.
1 skill
Matt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
15 skills · hooks
Persistent memory for Claude Code: the next session starts where the last one left off.
29 skills · 1 MCP server · hooks
Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications
5 skills · hooks
Cross-session memory and token savings for coding agents.
6 skills · 1 agent · 1 MCP server · hooks
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
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.