Skip to content

MVPavan/via / .claude/skills/skill-router

skill-router

Verified

Use when unsure which skill or command fits a task, when two seem to overlap, or when the user asks what skills or commands exist, what one does, or which to use.

Skill Router

Use the narrowest useful specialist contract. Clear bounded edits, ordinary explanations, factual lookups and routine verification often need no skill beyond AGENTS.md. This router is optional, not a startup checklist.

NeedRoute
Material unresolved directionbrainstorming
Durable plan/dependenciesplanning
Tracked multi-step implementationexecution
Substantive review or incoming feedbackcode-review
Explicit interviewgrilling; documented interview only when requested
Multi-session decision graphmanual wayfinder
Architecture improvement surveymanual improve-codebase-architecture
Describe a codebase or mechanismdirect source explanation, or manual codebase-research for a report
Intake-state decisionsmanual triage; routine tracking uses AGENTS.md/runtime context
Inline visualmanual show-me when requested; ordinary tables need no additional workflow

Legacy receiving-code-review, idea-refine, grill-me, grill-with-docs and teach-session are thin entrypoints to shared owners. Verification-before-completion and cost-estimate retain only manual migration guidance; do not recommend their retired workflows as automatic steps.

The catalog below is generated by the repository skill-catalog script. It lists canonical skills, not proof that every harness exposes them. Inspect actual runtime discovery when needed; root-only harness utilities and show-me may be absent from Codex links. Manual invocation syntax depends on the installed harness. An explicit user-supplied skill path can be read without changing integration files.

Model-invocable skills

authoring-for-agents, beads, brainstorming, code-review, codebase-design, design-evolve, document-review, domain-modeling, execution, first-principles-thinking, grilling, html-artifact, model-council, performance-optimization, perspective-council, planning, prototype, research, resolving-merge-conflicts, security, skill-router, systematic-debugging, test-driven-development

Slash-only workflows

SurfaceUse
/check-invariantsRun mechanically checkable invariants from the project overlay and report pass or fail.
/codebase-researchGraduated research on an external or unfamiliar codebase — survey its capabilities, map its architecture, or trace one mechanism — into durable reports under docs/research/codebases/.
/cost-estimateLegacy entrypoint; the uncalibrated code-volume and commit-based estimation method is retired.
/grill-meStart a decision interview using the shared grilling workflow.
/grill-with-docsStart a grilling interview and record authorized domain terminology and architectural decisions.
/i-have-adhdApply explicitly requested concise, easy-to-act-on communication preferences.
/idea-refineExplore a raw idea through brainstorming using the legacy idea-refine entrypoint.
/improve-codebase-architectureFind evidence-backed architecture improvements within a named area or requested codebase survey.
/phase-executionExecute one phase of a workstream roadmap.
/receiving-code-reviewUse when explicitly invoking the legacy receiving-code-review entrypoint; incoming feedback is owned by code-review.
/run-phasesRun every remaining phase of a workstream unattended.
/show-meShow the current topic visually, inline in the conversation.
/teachTeach a concept, walk through a session, or conduct an explicitly requested ongoing course.
/teach-sessionWalk through the current session using the shared teach workflow.
/triageEvaluate unevaluated beads issues and move each to an intake state — ready-for-agent, human, needs-info, backlog, or a wontfix close.
/verification-before-completionUse only when explicitly invoking this legacy entrypoint; verification policy is owned by AGENTS.md.
/wayfinderMap and resolve material decision dependencies across a multi-session effort.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 467kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 319kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 319kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 319kUpdated Claude CodeCodexVerifiedInstall