A workflow operating system for AI-assisted engineering. Task state, decisions, and plans live on disk as files, not in chat history, so context survives across sessions, tools, and restarts.
This repository also configures its own agents. See what fhorja.dev tells them →
Command Claude Code
Identify the invariants, non-goals, and forbidden changes for the active task, then persist them as INVARIANTSANDNONGOALS.md so implementation boundaries are locked before planning or coding. Use when impact analysis is done or mostly clear, the task may touch sensitive behavior or contracts or schema or runtime…
Command Claude Code
Snapshot the upstream Figma component library into docs/research/inventory/figmacomponents.md. For each Figma component: name, inferred tier, node ID, and whether a spec/code/story exists in the WOS-UI surface. Produces delta vs previous snapshot. Use after design ships a Figma library update or to seed the inventory…
Command Claude Code
Document a user journey from screen references and a user story input. Produces a journey spec doc matching the JOURNEYSPEC.md template with outcome, screen flow, components consumed, critical states, accessibility, security, and performance sections. Use when mapping a user flow across multiple screens. Do not use…
Command Claude Code
Read-only safety inspection of a third-party MCP server BEFORE it is added to a config or trusted. Reads the server's declared tool descriptions, input schemas, scopes, transport, and env/secret surface (not just its README), compares declared behavior against what the tools actually expose, scans for tool-description…
Command Claude Code
Document a reusable UX pattern (empty state, error handling, confirmation dialog, search-filter-sort, loading skeleton) that applies across multiple screens and projects. Produces a pattern spec doc matching the PATTERNSPEC.md template. Use when formalizing a recurring UX solution. Do not use when documenting a single…
Command Claude Code
Read-only cross-task board across every active task in all projects. Runs scripts/portfolio-review.sh to classify each active task as done-unclosed, blocked, my-move, stale, or in-flight, then renders one ranked board with a single recommended next action per row, so picking the next thread and catching…
Command Claude Code
Capture what a PR or team review changed about the intended behavior or approach, separate keep vs revert/replace, and produce the smallest safe set of updates to task memory and follow-on work. Does not implement product code. Use when a PR or team feedback requests a meaningful direction change (wrong fields, wrong…
Command Claude Code
Turn PR feedback (Greptile, CI, bots, humans) into a structured traceable backlog aligned with TASKSTATE.md, DECISIONS.md, and IMPLEMENTATIONPLAN.md so the next execution step can be a narrow implement-approved-slice or a small planning touch without losing alignment. Corrective scope only. Use when a PR is open or…
Command Claude Code
Prepare a clean delivery package for GitHub from the real git diff vs an explicit base branch, persisted as PRPACKAGE.md (or PRPACKAGE. .md for multi-repo tasks). Produces branch name, commit messages, fetch, checkout, add, commit, and push commands, PR title and body, and reviewer attention points; never invents work…
Command Claude Code
Run a short socratic intake that questions whether the stated problem is the right problem BEFORE a task exists, then write a task-level BRIEF.md (problem statement, success criteria, non-goals, recommended approach from 2-3 considered, named deliverables) that task-init consumes. Asks one question at a time, prefers…
Command Claude Code
Initialize a new project context inside the task repository before any task exists. Creates projects/ /, PROJECTCHARTER.md, REFERENCES.md skeleton, and active/ plus archive/ subfolders so subsequent task-init runs have grounded project-level memory to consume. Use when starting a brand-new project, product…
Command Claude Code
Shape the best possible prompt for the current task and workflow phase. Aligns the prompt content with the intended editor mode and the chosen command's required inputs, producing a copy-paste-ready handoff. Use when prompt precision would materially improve the next workflow action, the next step is known but the…
Command Claude Code
Design a pre-deploy release and rollout strategy for a change: pick the rollout pattern (feature flag, canary, blue-green, or full progressive delivery) by risk and infra, then specify the exposure ramp, the promotion metric and threshold that advances each step, and the rollback trigger and mechanism. Produces…
Command Claude Code needs its repo
Proactive defect-class detection that handles the lower-value half of code review (per Bacchelli and Bird 2013) so human reviewers stay focused on design, intent, and knowledge transfer. Catches convention drift, ordering bugs, type-safety gaps, security and multi-tenant invariants (CWE-grounded), and operability…
Command Claude Code
Turn unresolved or contradictory behavior, contract, or policy gaps into one canonical implementation-safe decision set, then persist the result in DECISIONS.md and TASKSTATE.md as explicit reviewable proposals (not silent intent changes). Use when key ambiguities have already been identified, the task has enough…
Command Claude Code
Resume the task from TASKSTATE.md and linked task artifacts, reconstruct the current truth, and determine the best next step. Used to bring back full task context after a session break or chat switch. Use when work is resuming in a new chat or session, the task was paused and needs fast reconstruction, or the user…
Command Claude Code
Review the current task changes for real correctness, safety, and maintainability risks before slice closure or PR prep, and recommend the smallest safe next step. Surfaces meaningful issues (not cosmetic feedback); not a replacement for external review systems. Returns no-op when the review would not materially…
Command Claude Code
Orchestrator-workers variant of screen-spec. Dispatches N Sonnet workers (1 screen each) in parallel to generate per-screen spec docs from Figma frames, then merges SCREENMAP.md and routes.md from per-worker partials. Use when screen count >= 6 AND Figma MCP is available AND foundations + SCREENMAP exist. Do not use…
Command Claude Code
Generate a screen specification from a Figma frame using MCP tools (getdesigncontext, getscreenshot). Produces a spec doc matching the SCREENSPEC.md template with layout sketch, components used, spacing observed, data dependencies, copy, accessibility notes, interactions, and error states. Use when documenting a…
Command Claude Code
Dedicated security review of the current task changes covering threat modeling, OWASP ASVS L1 checklist pass, auth/authz flow tracing, and dependency/secret scanning reminders. Distinct from review-hard (general risk) and repo-consistency-sweep (pattern matching). Activates when DECISIONS.md, IMPLEMENTATIONPLAN.md, or…
Command Claude Code
Take a draft artifact (IMPLEMENTATIONPLAN.md, SLICES/.md, or PRPACKAGE.md), run a structured critique against a locked per-artifact-type rubric, and produce a revised draft. Evaluator-optimizer pattern per ADR-0021. PROPOSED-by-default in Plan mode; user reviews critique AND revision before APPLIED. Distinct from…
Command Claude Code
Read-only safety inspection of a third-party agent skill or plugin BEFORE it is installed or trusted. Reads every file in the candidate directory (SKILL.md plus auxiliary and test files, not just the body), compares declared behavior against what the files actually do, scans for exfiltration, secret access, config or…
Command Claude Code
Decide whether the current slice is ready to close, distinguishing slice completion from full task completion, then persist the result in slice notes and TASKSTATE.md as explicit reviewable closure notes. For single-slice tasks may route directly to delivery. Returns no-op when slice memory would not materially…
Command Claude Code
Verify that the patterns the model is about to use for a given framework+version are current per official docs, and cache the result as CURRENTPATTERNS.md at the project level. Prevents the "gold-standard audit" anti-pattern where training-data defaults ship outdated patterns (e.g. Supabase getSession when getUser is…
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: