matshoppenbrouwers

security-liability-audit

Combined technical security and legal liability audit for code changes. Covers LLM/AI security, OWASP Top 10, agentic security (Lethal Trifecta), secrets detection, desktop app security, dependency supply chain, AND legal/liability risk (GDPR, EU AI Act, ToS/EULA coverage, data protection, consumer rights, cross-border transfers). Designed for solo developers in the EU serving a worldwide userbase. Use when: completing a feature implementation, before release, after any change that touches security-sensitive areas, AI features, data handling, payment logic, or user-facing terms. Triggers on "/security-liability-audit" or as a step in session-post-implementation.

session-add-task

Capture a new work item into the project's work root with its original request, interpretation and provenance. Use when something surfaces that should be done later — a bug, an optimization, a follow-up — and you want it recorded so it can be accepted and picked up. Produces a work item in the captured state, identity allocated by the runtime, and the regenerated sequence entry that shows it. Triggers on "/session-add-task" or when user says "add a task", "capture this for later", "put this on the backlog", or "remember to do X".

session-delegation

Dispatch named tasks of one work item to agents. Loads each task record through the runtime, resolves the permitted prerequisite closure, and records each result against its claimed identity. Use after /session-task-planning has written task records. Triggers on "/session-delegation" or when user says "execute the tasks", "run the plan", or "dispatch the agents".

session-gatekeeper

Triage incoming work (GitHub issues, feature requests, surfaced ideas) and route it — trivial aligned fixes go into the task sequence, while significant or direction-divergent items are escalated to a cowork research-design session. Grounds every decision in the app's architecture and product direction (PRD.md). Use as the front-of-chain intake funnel. Triggers on "/session-gatekeeper" or when user says "triage these issues", "process the backlog of issues", "should we do this issue", or "is this worth a research-design session".

session-groom

Groom the backlog by researching captured work items that carry no verified breakdown and recording one against each. Use to keep the work root healthy — every item researched, verified, and ready to be accepted — especially for items captured as raw one-liners. Safe to run periodically via /loop. Triggers on "/session-groom" or when user says "groom the backlog", "prepare the sequence", "fill in the task breakdowns", or "tidy up SEQUENCE.md".

session-init

Bootstrap the project documentation structure for session-flow skills. Creates the work root with its namespace, plus research/, plans/, testing/, and architecture/ directories. Run once when adopting session-flow in a new project. Triggers on "/session-init" or when user says "initialize project", "setup docs structure", or "bootstrap project".

session-next

Implement the next work item from the project's backlog. Asks the runtime to select one accepted item, claims it, executes one bounded unit, and records the result against that identity. Use when the user wants to make progress on the backlog without naming a specific item. Triggers on "/session-next" or when user says "implement the next task", "do the next task", "what's next — just do it", or "work the backlog".

session-post-implementation

Post-implementation refinement workflow. Use after completing a major feature or plan implementation to simplify, review, and document the code. Triggers on "/session-post-implementation" or when user says "run the iteration workflow" or "polish this implementation".

session-release

End-to-end release workflow that bumps version, packages artifacts, and verifies satellite content (docs sites, marketing sites, changelogs) is up-to-date. Combines project-specific release skills with a generic release surface scan. Triggers on "/session-release" or when user says "prepare the release", "do a release", or "release workflow".

session-research-design

Deep research and design workflow for complex features. Use before /session-task-planning when a feature requires exploring implementation approaches, analyzing existing code, and designing an architecture before breaking into tasks. Attaches to a work item by identity and produces a research report plus the spec and plan that work warrants. Triggers on "/session-research-design" or when user says "research this", "explore approaches for", or "design this feature".

session-task-planning

Break an accepted work item into Claude Code-scoped task records with parallelization analysis. Use when breaking down a multi-step implementation into discrete tasks that can each be completed in one Claude Code session. Produces one task record per task under the item's identity, each carrying its dependencies, write boundary, and criteria. Triggers on "/session-task-planning" or when user says "break this into tasks", "plan the tasks", or "create a task list".

session-verify

Evidence-based verification workflow for completed features and plans. Produces a falsifiable proof artifact that implementation matches design spec, identifying defects, stubs, and spec-vs-reality gaps. Triggers on "/session-verify" or when user says "verify this feature", "prove this works", "run verification", or "confirm the implementation".

update-architecture

Surgical updates and current-state initialization for the project's architecture documentation root. Ensures token-efficient edits, current data-flow and system-structure notes, architecture index hygiene, and doc splitting checks. Called by session-post-implementation when architecture docs are in scope. Triggers on "/update-architecture" or when user says "update arch docs", "sync architecture", "map current architecture", "initialize architecture docs", or "docs are out of date".

session-brainstorm

Use when the user says "let's brainstorm" or "let's think about", or brings an idea that is not yet a task. Explores intent and options in conversation, ends with a short design the user approves, captures it as a work item, then routes to implementation or to /session-research-design.

session-debug

Use when the user reports a bug, a failing test, or unexpected behaviour and wants it diagnosed before it is fixed. Finds the root cause, tests one hypothesis at a time, and stops to question the design after three failed fixes.

session-repair

Transform a legacy layout into the work root, or reconcile a work root that has drifted. Surveys read-only, refuses unsafe ground, prints the per-item plan, applies only on explicit confirmation, then verifies the result entry by entry. Triggers on "/session-repair" or when user says "migrate the backlog to the work root", "the sequence and the records disagree", "reconcile the work root", or "repair session-flow state".