By simon-potter
14 specialist subagents covering every phase of the SPWorkflow engineering cycle
Approve plan agent. Reads the task list generated by spec, assesses quality (atomicity, testability, clarity, ordering), flags problems, and presents the plan for human approval before building starts. The task list was created by the specifier — this agent reviews it, not produces it. Use after spec has run.
Phase 3 build agent. Reads the current task from the OpenSpec task list, implements it, marks it complete, and stops. Does not move to the next task. Implements via opsx:apply. Use for the main build loop, one task at a time.
Capture agent. Accepts any input — issue tracker ticket (YouTrack default; Jira and others supported), Slack message, file, or freeform description — classifies it as a bug or a change, then runs the appropriate path. Bug path runs systematic root-cause investigation, classifies fix complexity (content/config vs code), and produces todo/BUG-{slug}.md. Change path runs a lightweight qualification check and produces todo/{slug}.md. Prompts to create a tracker ticket if none exists. Delegates to spwf:capture.
Gate — Challenge agent. Reads the ideation file and interviews relentlessly until all open questions are resolved. One question per message. Does not proceed to spec until gaps are closed.
Analyses PHP code and flags bad practices with suggested fixes, grouped by risk category. Framework-aware (Laravel, Symfony, WordPress). Read-only — flags issues and proposes fixes, never edits files. Use before merging PHP changes, when reviewing a PR touching backend logic, or when auditing a legacy PHP codebase. Complement to security-scan (OWASP/injection depth) and php-code-simplifier (applies safe refactors).
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Simon's engineering workflow, packaged as two installable Claude Code plugins.
[status] → [Capture] → Challenge → Spec → Approve plan → Build → Simplify → PR Create → PR Review → Close
(orient) (pre) (gate) (1) (2) (3) (4) (5) (6) (post)
| Step | Command | Invokes | Why | Produces |
|---|---|---|---|---|
| Orient | /spwf:wfstatus | — | Start of session: where am I, what's incomplete, what's next — heuristics across git state, OpenSpec changes, and todo backlog | Dashboard + suggested next action |
| Capture | /spwf:capture [source] | Issue tracker MCP (YouTrack default; Jira supported) | Accepts a tracker ticket (e.g. ACAD-42), file, or freeform description; classifies as bug or change automatically. Bug path: systematic root-cause investigation → hypothesis. Change path: lightweight qualification, one question at a time. | todo/{slug}.md or todo/BUG-{slug}.md |
| Challenge | /spwf:challenge todo/{slug}.md | — | Interviews relentlessly until all open questions are resolved, then runs a scope-sizing check — recommends splitting into independent changes if the work spans natural boundaries, or proceeds as one change if tightly coupled | Resolved ideation file; or N child todo files + original marked status: split |
| Spec | /spwf:spec todo/{slug}.md | openspec CLI | Formalises the challenged idea into a structured spec | openspec/changes/{id}/proposal.md, design.md, tasks.md, specs/ |
| Approve plan | /spwf:approve-plan | — | Quality check (blocking) + adversarial review via Skeptic/Architect/Minimalist lenses (advisory); explicit human go/no-go before building | Approved task list or flagged issues to resolve |
| Build | /spwf:build | write-tests → opsx:apply → run-tests → debug-recovery → opsx:verify | Red-Green-Verify per task, loops until all done; spec sign-off after all tasks complete | All tasks complete, tests green, spec aligned |
| Simplify (TDD Refactor) | /spwf:simplify | — | Clean up the implementation with tests as a safety net; flags judgment calls | Cleaner diff; flag list |
| PR / MR Create | /spwf:pr-create | dep-audit · forge CLI (glab default; gh supported) | Pre-flight checks (gitleaks, semgrep, dep-audit across all ecosystems + Docker) then request creation via the forge auto-detected from git remote; CI/CD owns the rest | PR / MR URL |
| PR / MR Review | /spwf:pr-review <ref> | forge CLI (glab mr view/diff default; gh pr view/diff supported) | Structured review before merge; catches regressions and drift | Review report with verdict |
| Close | /spwf:close [todo/{slug}.md] | retrospective → opsx:archive → Issue tracker MCP | Final phase — runs the full retrospective (learn-from-mistakes, spec audit, doc-lint, workflow-lint, optional changelog), then after explicit confirmation marks the todo file complete, archives the OpenSpec change, and transitions the linked tracker ticket to its done state (per .spwf/tracker.yaml) | Closed todo, archived change, tracker ticket marked done |
A second class of skills sits outside the main workflow. These are cross-cutting maintenance tools — run them between sessions, on a cadence, or when something feels off. They don't produce code; they keep the workspace itself in good shape.
npx claudepluginhub simon-potter/spwf --plugin spwf-agents28 workflow skills covering the full engineering cycle: capture, challenge, spec, plan, build, test, review, simplify, ship, learn, and quality maintenance
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Comprehensive PR review agents specializing in comments, tests, error handling, type design, code quality, and code simplification
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Use this agent when creating user interfaces, designing components, building design systems, or improving visual aesthetics. This agent specializes in creating beautiful, functional interfaces that can be implemented quickly within 6-day sprints. Examples:\n\n<example>\nContext: Starting a new app or feature design