/workflow

name: workflow
description: "Structured work: multi-phase tasks, feature builds, planning, objective loops, hill climbing."
user-invocable: true
context: fork
allowed-tools:
  - Read
  - Write
  - Edit
  - Bash
  - Glob
  - Grep
  - Skill
  - Agent
  - Task
routing:
  force_route: true
  not_for: "code review (use review), testing (use testing), security (use security)"
  triggers:
    - "workflow"
    - "multi-phase task"
    - "feature design"
    - "feature plan"
    - "feature implement"
    - "build feature end to end"
    - "full feature lifecycle"
    - "write spec"
    - "define requirements"
    - "create plan"
    - "create tasks"
    - "keep working until"
    - "iterate until done"
    - "drive this to done"
    - "make this faster"
    - "speed this up"
    - "reduce latency"
    - "profile and optimize"
    - "hill climb on this metric"
    - "tidy up"
    - "clean up"
    - "untangle"
    - "reorganize"
    - "structured pipeline"
    - "phased execution"
  category: process
  pairs_with:
    - review
    - testing
    - security
    - pr-workflow

Workflow Skill

Five modes for structured multi-phase work. Match the request, follow that mode's instructions.

Mode Selection

| Request pattern | Mode | |---|---| | Feature design/plan/implement/validate/release, end-to-end | [Feature Lifecycle](#feature-lifecycle) | | Write spec, define requirements, create plan, interview, pause/resume | [Planning](#planning) | | Keep working until, iterate until done, drive to done | [Objective Loop](#objective-loop) | | Make faster, speed up, reduce latency, profile, hill climb | [Hill Climb](#hill-climb) | | All other structured workflows: review, debug, refactor, research, create, explore, upgrade | [Ad-Hoc Workflow](#ad-hoc-workflow) |

---

Feature Lifecycle

Phase-gated workflow: DESIGN > PLAN > IMPLEMENT > VALIDATE > RELEASE. Each phase must pass its gate before the next begins.

Phase Routing

If `.feature/` exists, check state: `python3 ~/.claude/scripts/feature-state.py status`. Route to the indicated phase.

If no feature state exists, determine entry from intent:

• "design", "think through", "explore approaches" -> DESIGN
• "plan", "break down", "create tasks" -> PLAN (requires completed design)
• "implement", "execute plan" -> IMPLEMENT (requires completed plan)
• "validate", "quality gates" -> VALIDATE (requires completed implementation)
• "release", "merge", "ship it" -> RELEASE (requires passed validation)
• "end to end", "full lifecycle" -> DESIGN (start from beginning)

Phase References

Load the phase reference, then follow it exactly:

| Phase | Reference | Produces | |---|---|---| | DESIGN | `references/fl-design.md` | design.md | | PLAN | `references/fl-plan.md` | Wave-ordered task list | | IMPLEMENT | `references/fl-implement.md` | Code changes | | VALIDATE | `references/fl-validate.md` | Quality gate report | | RELEASE | `references/fl-release.md` | Merged PR | | End-to-end | `references/fl-pipeline.md` | Full lifecycle | | State conventions | `references/fl-shared.md` | -- | | Error recovery | `references/fl-error-handling.md` | -- |

State operations use `python3 ~/.claude/scripts/feature-state.py` only. Never manipulate state files directly.

---

Planning

Spec writing, plan creation, interviews, ambiguity triage, and session pause/resume. Planning owns specs and saved plans; execution goes through the process skill's Subagent mode or workflow dispatch.

Sub-mode Routing

| Signal | Reference | |---|---| | Write spec, user stories, define requirements, scope, acceptance criteria | `references/pl-spec.md` | | Discuss ambiguities, resolve gray areas, pre-planning discussion | `references/pl-pre-plan.md` | | Interview me, depth-first review, "not sure", "where do I start", "poke holes" | `references/pl-depth-first-interview.md` | | Implicit ambiguity or unclear implementation choices | `references/pl-ambiguity-triage.md` | | Another person holds needed facts or approval | `references/pl-human-source-elicitation.md` | | Observation can settle a disputed choice | `references/pl-empirical-prototype.md` | | Phase or session transition near | `references/pl-context-boundary.md` | | Create plan, task plan, file-backed planning | `references/pl-plan-files.md` | | Check plan, validate plan, pre-execution check | `references/pl-check.md` | | List plans, show plan, complete plan, manage plans | `references/pl-manage.md` | | Pause, save progress, handoff, stopping for now | `references/pl-pause.md` | | Resume, continue, pick up where I left off | `references/pl-resume.md` |

For interviews, batch independent questions into frontier rounds. Ask dependent questions sequentially. Include a recommendation per question.

---

Objective Loop

Iterate-until-verified-done loop. A user states an objective with verifiable done-criteria; each iteration routes one /do cycle, verifies by executing the criteria, and reschedules until verified-done or budget-stop.

Phase 1: SPEC

Gather from the request: objective statement, DONE-CRITERIA (verifiable checks), iteration budget (default 5), NOT-DONE-YET guardrails (what may never be done to satisfy a criterion).

DONE-CRITERIA types: `command` (preferred -- deterministic command with expected exit code/output) or `rubric` (only when no mechanical check exists -- frozen at SPEC time, graded by a fresh-context agent).

Phase 2: STATE

Write `.objective/<slug>/state.md` from `references/ol-state-file.md`. Wakeups resume from the state file, never conversation memory.

Phase 3: ITERATE

Plan the smallest next step. Route through /do: classify -> route -> dispatch agents -> evaluate. The loop dispatches exclusively through /do -- never edit inline.

Phase 4: VERIFY

Run every done-criterion check. A worker's "passes" claim never substitutes for re-running.

• `command`: run it, paste exit code and output into iteration log.
• `rubric`: dispatch fresh-context sub-agent (did NOT produce the work) with