skills/handoff-iteration-loop/SKILL.md
Use when coordinating multi-iteration work with explicit `_working/{TOPIC}/` handoffs, binary sprint contracts, and independent reviewer gating after implementation.
npx skillsauth add michael-f-bryan/skills handoff-iteration-loopInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Multi-agent pattern inspired by Harness design for long-running application development (Anthropic): separate generation from evaluation, structured artefacts across iterations, sceptical external QA, contracts before heavy work, simplify the harness when the task no longer needs every role — and real dev-team shape: discovery / Q&A → goals → sprints → loop, with Git commits as the implementor goes.
| If you are… | Read first | Then if needed |
| ----------------------------------------------------------------------- | --------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------- |
| Spinning up the workflow | This file + workflow.md | discovery.md if the ask is vague; then role files |
| Discovery / customer goals | discovery.md | workflow.md §Phases |
| Redesigning the harness (adding/removing roles, changing contracts) | harness-principles.md + workflow.md | Role files only if prompts change |
| Coordinator (main agent orchestrating Task / sub-agents) | coordinator.md | workflow.md §Contracts, any single planner.md / reviewer.md for prompt patterns |
| Planner sub-agent | planner.md | workflow.md §Handoff filenames + §Sprint contract |
| Implementor sub-agent | implementor.md | workflow.md §Contracts, planner.md §Anti-patterns (mirror only outcomes) |
| Reviewer sub-agent | reviewer.md | workflow.md §Sprint contract + §Evidence |
Do not load all role files into one context unless the coordinator is authoring a full runbook in one shot.
(Optional) Discovery → _working/{TOPIC}/GOALS.md [see discovery.md]
↓
Coordinator → Planner writes _working/{TOPIC}/iter{N}.1-planner.md
→ Implementor (git commits as you go + iter{N}.2-implementor.md)
→ Reviewer (verification + iter{N}.3-reviewer.md)
→ Coordinator gates iter{N+1} or stops; updates project HANDOVER/STATE line
Markdown handoffs + GOALS.md carry intent; git carries implementation history. No chat-memory reliance.
_working/{TOPIC}/ (see workflow.md §Tiering):
iter{N}.1-planner.md, iter{N}.2-implementor.md, iter{N}.3-reviewer.mditer{N}.2-implementor.md, iter{N}.3-reviewer.md (+ reference the active planner contract)iter{N}.2-implementor.md (+ explicit “risk accepted” line)## Sprint contract (binary checks) — reviewer executes it mechanically (workflow.md).MUL-55 Grafana temp dashboard: _working/mul55-dash/iter6.1-planner.md (contract + protocol), _working/mul55-dash/iter6.3-reviewer.md (sceptical PASS table), _working/HANDOVER-mul55-cost-dashboard.md (index line pattern).
documentation
Use when handling multi-step tasks, investigations, or long sessions where working notes, interim findings, and scratch planning are needed to keep context and handoffs clear.
documentation
When generating or editing markdown content, actively look for existing pages to link to and incorporate relevant wikilinks so content is interconnected. Use when writing notes, docs, or any .md content.
development
Wait for a GitHub Actions run to finish with minimal terminal output and a reliable exit code. Use when an agent must wait for CI to pass (e.g. after push, after opening a PR, or when verifying a specific run). Prefer gh run watch with --exit-status and --compact to avoid flooding context with poll output.
testing
Use when drafting, editing, or evaluating communication that should sound like Michael across public technical prose, internal notes, agent prompts, engineering feedback, and concise operational updates.