Skills Catalog

RaiSE ships with 81 public skills organized by lifecycle (plus 0 internal skills not listed). Each skill is a structured workflow defined in a SKILL.md file that your AI partner follows step by step.

Invoke any skill by typing its name (e.g., /rai-story-design) in your AI assistant. See Skills concept for how skills work, or Create a Skill to build your own.

Epic Lifecycle

Epics are bodies of work spanning 3–10 stories. See Your First Story for context.

Skill	Description
/rai-epic-close	Default variant: retrospectiva + tag + fixVersion. 3 steps vs 8 enterprise. Transitions engine-owned (RAISE-15034).
/rai-epic-close-enterprise	Close epic with retrospective, push, and merge request. Use after all stories done.
/rai-epic-design	Default variant: graph query + code read + story breakdown + ADR-135 artifacts. Produces scope.md + design.md with Machine section.
/rai-epic-design-enterprise	Design epic scope, stories, and architecture. Use for work spanning 3-10 features.
/rai-epic-docs	Generate architecture docs and publish to Confluence. Use before epic close.
/rai-epic-journal	Write a narrative epic journal as a self-contained HTML file. Use after epic stories are done.
/rai-epic-plan	Sequence epic stories into milestones and dependencies. Use after epic design, and consume epic UX design when present.
/rai-epic-predesign-evidence	PDCV evidence gathering: gemba walk + backlog scan → predesign.json + capability-map.md. Runs before epic design to prevent reimplementation.
/rai-epic-start	Default variant: git status + raise_docs_write brief/scope + backlog. 3 steps vs 9 enterprise.
/rai-epic-start-enterprise	Initialize epic directory, brief, and tracker entry. Use to begin a new epic.
/rai-epic-ux-design	Define workflows, surfaces, and trust boundaries for user-facing epics before planning. Use after epic design when interaction design matters.
/rai-mr-create	Create a governed GitLab MR or GitHub PR through provider-specific admission.
/rai-mr-merge	Bring an open merge request up to date with its target, wait for CI to go green, and merge it through the SCM adapter. The single point where a branch lands.

Story Lifecycle

Stories are the unit of deliverable work. Every story follows the same six phases.

Skill	Description
/rai-architecture-review	Evaluate design proportionality using Beck's four rules. Use after implementation.
/rai-quality-review	Audit code for semantic bugs, type lies, and test muda. Use after implementation.
/rai-story-close	Canonical story close: gates + raise_story_close_full + scope.md. Sin señales de observabilidad (raise_story_close_full las emite server-side). ADR-143 B2.
/rai-story-close-enterprise	Merge story branch to dev and update tracking. Use after story review.
/rai-story-design	Create lean story spec for human review and AI alignment. Use before story plan.
/rai-story-implement	Execute plan tasks with TDD and validation gates. Use after story plan.
/rai-story-plan	Decompose story into atomic tasks with TDD verification. Use after story design.
/rai-story-review	Extract learnings and persist patterns from completed story. Use after implementation.
/rai-story-start	Canonical story start: fetch+merge + raise_story_open. Sin señales de observabilidad (raise_story_open las emite server-side). ADR-143 B1.
/rai-story-start-enterprise	Create story branch and scope commit. Use to begin story work.

Bugfix Lifecycle

Bugs follow a structured pipeline with root cause analysis and post-incident review. See Bugfix Lifecycle.

Skill	Description
/rai-bugfix-analyse	Root cause analysis using the method best suited to the bug. Phase 3 of bugfix pipeline.
/rai-bugfix-close	Default variant: push + MR + Jira sync. 12 steps vs 26 enterprise. Assumes pipeline-orchestrated.
/rai-bugfix-close-enterprise	Push branch, create MR, verify artifacts complete. Phase 7 of bugfix pipeline.
/rai-bugfix-fix	Default variant: TDD + gates + quality-review. 11 steps vs 19 enterprise, no HITL per-task.
/rai-bugfix-fix-enterprise	Execute fix tasks with TDD and all validation gates. Phase 5 of bugfix pipeline.
/rai-bugfix-pir	Post-Incident Review — learning narrative and improvement actions. Phase 8.5 of bugfix pipeline.
/rai-bugfix-plan	Decompose fix into atomic TDD tasks. Phase 4 of bugfix pipeline.
/rai-bugfix-review	Retrospective, pattern extraction, and process improvement. Phase 6 of bugfix pipeline.
/rai-bugfix-start	Default variant: branch + reproduce + scope. 11 steps vs 25 enterprise.
/rai-bugfix-start-enterprise	Initialize bug branch, reproduce, and create scope artifact. Phase 1 of bugfix pipeline.
/rai-bugfix-triage	Classify bug in 4 dimensions and set Jira custom fields. Phase 2 of bugfix pipeline.

Bugfix Lifecycle (Lean)

Skill	Description
/rai-bugfix-close-lean	Lean residue de rai-bugfix-close: push + MR + Jira sync. 10 pasos vs 26 enterprise. Asume pipeline-orchestrated.
/rai-bugfix-fix-lean	Lean residue de rai-bugfix-fix: TDD + gates + quality-review. 11 pasos vs 19 enterprise, sin HITL per-task.
/rai-bugfix-start-lean	Lean residue de rai-bugfix-start: branch + reproduce + scope. 11 pasos vs 25 enterprise.

Session Lifecycle

Sessions are the daily work container. They provide continuity across conversations.

Skill	Description
/rai-session-close	Capture session outcomes and update memory. Use to close a working session.
/rai-session-diary	Write a narrative session diary entry to Confluence. Use before session close.
/rai-session-start	Load context and propose session focus. Use at the start of every working session.

Spike Lifecycle

Skill	Description
/rai-spike-close	Capture spike decision and close with minimal retrospective. Use after research or prototype complete.
/rai-spike-journal	Write a narrative spike journal entry for the team on Confluence. Use after a spike's decision is reached, before close.
/rai-spike-start	Initialize a time-boxed spike with scope artifact. Use to begin research or prototype work.

Initiative Lifecycle

Skill	Description
/rai-initiative-delivering	Default variant: validate-only — verifies epic decomposition without creating artifacts. 3 steps vs 5 enterprise.
/rai-initiative-shaped	Default variant: hypothesis → treatment.md + mvp-spec.md with Machine section (ADR-135). 4 steps vs 8 enterprise.

Creation

Skill	Description
/rai-kc-build	Build a Knowledge Cartridge from source documents — guided flow from corpus analysis to extraction and review.

Discovery

Skill	Description
/rai-discover	Scan codebase, extract symbols, and build knowledge graph. Use for codebase discovery.

Meta

Skill	Description
/rai-skill-create	Create a new skill through guided conversation. Use to add skills to the framework.
/rai-skillset-manage	Manage skill sets via CLI. Use to create, compare, or customize skill sets.

Utility

Skill	Description
/rai-adversarial-review	Bounded single-shot adversarial red-team of an analysis or design result — fresh context attacks the findings, verifies evidence against code/state, returns holds/weak/wrong + what's missing. Use after any research/analysis/design phase, before acting on it.
/rai-backlog-setup	Configure a Jira instance for all RaiSE skills in one session: connection, custom fields, workflow statuses, and link types — backlog.yaml complete.
/rai-bug-file	File a bug with diagnostic precision — repro, classification, root cause hypothesis, and Jira issue creation. Intake quality gate for the bugfix pipeline.
/rai-cartridge-agent	Build a Knowledge Cartridge using the local agent's own inference — no external LLM API key. Propose schema, extract nodes, and relate them via emit-work/ingest.
/rai-corpus-curate	Discover and curate a project's documents into a governance corpus — classify by relevance, drop noise, confirm with the developer before extraction.
/rai-debug	Find root cause using 5 Whys, Ishikawa, Gemba. Use when encountering errors or defects.
/rai-docs-setup	Configure the docs adapter for all RaiSE skills in one session — docs adapter config complete. Currently supports Confluence.
/rai-docs-setup-advanced	Configure advanced custom routing — add new Confluence spaces and set per-artifact parent_path hierarchy. Full wizard for complex multi-space orgs.
/rai-docs-update	Sync module docs with knowledge graph. Use when architecture docs drift.
/rai-doctor	Run self-diagnostics and guide fixes. Use to troubleshoot RaiSE issues.
/rai-mcp-add	Register an MCP server conversationally. Use to add a new MCP integration.
/rai-mcp-remove	Remove an MCP server with dependency checking. Use to unregister a server.
/rai-mcp-status	Check health of registered MCP servers. Use to diagnose MCP connectivity.
/rai-onboard-repo	Connect a repository to the RaiSE server. Detects what's done, guides through what's missing.
/rai-pipeline-run	Thin pipeline orchestrator. Dispatches phases as forked agents, never reads work files.
/rai-problem-shape	Shape a vague idea into a problem brief. Use before epic design for new initiatives.
/rai-project-create	Set up governance for a new project. Use after rai init on a greenfield project.
/rai-project-onboard	Discover conventions and set up governance. Use after rai init --detect on existing code.
/rai-project-upgrade	Upgrade a RaiSE project from an older version (e.g. v2.4) to the current version. Detects version, scaffolds missing structure, migrates data.
/rai-research	Research pipeline entry point. Routes to /rai-research-run (adaptive pipeline with depth-based phase selection).
/rai-welcome	Onboard a developer to RaiSE interactively. Use for first-time setup.
/rai-worktree-close	Close a git worktree — mark it closed in DB and suggest /rai-mr-create if work is ready to ship. Works with or without a registered worktree.
/rai-worktree-open	Open a new git worktree with full environment setup. Use to start parallel story work.

Other

Skill	Description
/rai-pipeline-run-lean	Thin pipeline orchestrator. Dispatches phases as forked agents, never reads work files.
/rai-research-critic	Completeness critic — identifies coverage gaps and triggers re-search if needed. Deep depth only.
/rai-research-evidence	Merge, deduplicate, and catalog findings with evidence levels. Skipped for quick depth.
/rai-research-frame	Define research question, determine depth route, create work directory.
/rai-research-run	Adaptive research pipeline orchestrator. Dispatches phases as forked agents, routes by depth.
/rai-research-scope	Decompose research question into 3-5 divergent search angles. Skipped for quick depth.
/rai-research-search	Execute searches across angles using Tavily, Brave, ddgr, and WebFetch. Scales by depth.
/rai-research-synthesize	Produce final research report with recommendations, confidence levels, and governance linkage.
/rai-research-verify	Adversarial verification of major claims using independent forks. Deep depth only.