AI Agent Workflow Collection

Patterns for adding, renaming, removing, or retyping fields in DorkOS user config. Use when editing UserConfigSchema, MarketplacesFileSchema, or any conf-backed store — walks the Zod field → defaults → conf migration → docs → test lifecycle end-to-end.

Regenerate DorkOS product screenshots and video loops from the real UI, and manage the shot registry, human overrides, and version archives that feed the marketing site, docs, and changelogs. Use when marketing/docs media is stale after UI changes, a new feature or docs page needs a money shot, a person wants to override an automated capture, or a release calls for a fresh capture or archive — anything under apps/site/public/product/ or the apps/e2e/capture pipeline.

The /flow engine's CAPTURE stage — quick, low-commitment intake of a new piece of work into the tracker as an idea, without doing full triage or evaluation. Use when the goal is to get a thought captured cleanly so it survives, not to assess, classify deeply, or plan it. Generalizes the legacy /linear:idea capture flow (retired in spec 257); PM-agnostic.

Analyzes user prompts for gaps, ambiguities, and unstated assumptions, then asks clarifying questions before work begins. Use when requests are vague, lack acceptance criteria, or have hidden complexity.

The /flow engine's DONE stage — report completion on a work item, move it to Done with the agent/completed label, create any follow-up work, run a project pulse check for the next loop action, and clean up the worktree. Use when running /flow:done or advancing an approved work item into the DONE stage.

Guides systematic debugging methodology and troubleshooting approaches. Use when investigating bugs, tracing issues, or teaching debugging patterns.

Systematically investigates failing tests, distinguishes between test bugs and implementation bugs, and drives a fix with verification. Use when the user wants to debug failing tests.

Systematically investigates TypeScript compiler errors, traces the real type mismatch, and applies a minimal verified fix. Use when the user wants to debug TypeScript type errors.

Documentation organization, maintenance, and cleanup. USE WHEN: organizing docs, cleaning project root, updating documentation, checking for redundancy, maintaining docs structure. NOT FOR: technical implementation (use relevant technical skill).

First-touch triage for incoming GitHub issues on dork-labs/dorkos. Use when DorkBot is asked to triage the issue queue: label each new issue by type and runtime, find likely duplicates, and ask the reporter for missing reproduction steps. Read-and-suggest by default; only comments or edits labels when explicitly allowed to act.

Orchestrates parallel implementation of decomposed specifications with incremental progress tracking. Use when running the /flow:execute stage (the EXECUTE stage of the /flow engine).

Turns a feature brief, rough notes, or an existing partial design into a structured ideation artifact and next-step plan. Use when the user wants to explore a feature or shape rough requirements into a structured ideation artifact.

Implements a validated specification, keeps execution progress durable, and advances linked project work cleanly. Use when the user wants to implement or execute a validated specification.

The /flow engine's tracker adapter — the single skill that owns EVERY Linear MCP / Composio call and normalizes Linear into the generic WorkItem shape. Use whenever a /flow stage skill or the loop engine needs to read or write the tracker (claim, transition, comment, inbox, relations, evidence, sub-issues). All flow tracker I/O routes through here; no other flow skill or command may touch a tracker string.

Keeps the Dev Playground current with the application. Use when editing UI components in apps/client/src/ — assesses playground candidacy, checks existing playground coverage, and guides adding or updating showcases. Also use when directly editing files in apps/client/src/dev/, building new widgets or features, or when the user mentions the dev playground, component showcase, or visual testing. Proactively evaluate whether edited components belong in the playground even if the user doesn't ask.

Canonical rules for spec manifest management — statuses, transitions, and the manifest-ops script. Auto-loads when working with spec files.

Orchestrates parallel execution of AI agents with dependency analysis and batch scheduling. Use when coordinating multiple concurrent tasks, optimizing task ordering, or when multiple independent agents would benefit the workflow.

Guides organization of code using Feature-Sliced Design (FSD) architecture. Use when structuring projects, creating new features, determining file and layer placement, or reviewing architectural decisions. Also monitors codebase size and proactively suggests structural improvements.

Resolves DorkOS session URLs to JSONL transcript files on disk and reads them. Use when the user shares a session URL, mentions "read this transcript", "read this chat", or references a DorkOS session they want analyzed.

The /flow engine's SPECIFY stage — turns a validated ideation artifact into an implementation-ready specification, resolves its open decisions, and seeds draft ADRs. Use when work is ready to move from IDEATE to a frozen spec. PM-agnostic; all tracker I/O routes through the linear-adapter skill.

Implements the Calm Tech design system using Tailwind CSS v4 and Shadcn UI. Use when writing styles, building components, or theming. For design decisions, see designing-frontend.

Keeps skill definitions aligned between Claude Code and Codex. Use when creating, migrating, renaming, or updating skills that must work across .claude/skills and .agents/skills, or when auditing drift between the two systems.

The /flow engine's agent-as-team-member loop — the agent participates in a shared tracker like a teammate. Each tick it polls its inbox (assigned-to-me + @mentions + new comments), decides respond/act/ignore per the five comment-response rules, claims work with durable agent/ labels, asks the human via comment+assign when genuinely stuck (soft-escalation), and writes answers back as durable memory so the same question is never asked twice. Use when the agent is operating continuously inside a shared tracker. PM-agnostic; all tracker I/O routes through the linear-adapter skill.

The /flow engine's TRIAGE stage — classify and route incoming work. Classifies a freeform brief/idea/bug/question into the right work-item type, or evaluates an already-captured item (accept/reject/needs-research/needs-refinement) and routes it simple-vs-complex (stay-in-tracker task vs escalate to the spec workflow). Use when the goal is to evaluate and route work, not just capture it. Generalizes the legacy /pm triage/intake path; PM-agnostic.