JPeetz/agent-skills / skills/development/mattpocock-dev-workflows
mattpocock-dev-workflows
VerifiedUse when you need Matt Pocock's engineering workflow patterns: spec synthesis, deep-module design, domain modeling, two-axis code review, or disciplined bug diagnosis. A meta-skill covering five complementary sub-patterns for real engineering — not vibe coding.
| File | Kind | Size |
|---|---|---|
| CHANGELOG.md | markdown | 2.0 KB |
| LICENSE | other | 1.1 KB |
| SKILL.md | skill md | 20.6 KB |
| evals/evals.json | config | 5.2 KB |
| references/adr-format.md | markdown | 2.7 KB |
| references/code-review-smell-baseline.md | markdown | 2.2 KB |
| references/codebase-design-glossary.md | markdown | 2.7 KB |
| references/context-format.md | markdown | 2.3 KB |
| references/deepening.md | markdown | 2.5 KB |
| references/design-it-twice.md | markdown | 2.6 KB |
| references/diagnosing-bugs-phases.md | markdown | 7.2 KB |
| references/to-spec-template.md | markdown | 1.6 KB |
evals/evals.json
View on GitHub{
"schema_version": "1.0",
"evals": [
{
"name": "to-spec-synthesis-no-interview",
"description": "Verify to-spec synthesizes conversation without interviewing the user",
"test_input": "User has just spent 20 minutes discussing a new payment feature. They say: 'Write this up as a spec.' The conversation contains: an actor (mobile bank customer), a feature (view transaction history with filters), and a benefit (better spending decisions).",
"expected_output": "A spec containing: Problem Statement, Solution, User Stories (including 'As a mobile bank customer, I want to view my transaction history with filters, so that I can make better informed spending decisions'), Implementation Decisions (seams, modules), Testing Decisions, Out of Scope. No interview questions.",
"assertions": [
"Does NOT contain any question asking the user for clarification (e.g. 'What actor?' 'What benefit?')",
"Contains 'User Stories' section with at least 3 user stories",
"Contains 'Implementation Decisions' section",
"Contains 'Testing Decisions' section",
"Contains 'Out of Scope' section",
"Understands context from the existing conversation without re-asking"
]
},
{
"name": "codebase-design-deep-module-identification",
"description": "Verify codebase-design correctly identifies deep vs shallow modules",
"test_input": "Given a module: function processOrder(order, gateway, logger, notifier) { ... } with 50 lines of implementation that just delegates to 4 sub-functions, and a module: function validateOrder(order) { if (!order.id) throw Error('missing id'); if (!order.email) throw Error('missing email'); return true; } with 10 lines of implementation and 1 parameter.",
"expected_output": "processOrder should be identified as shallow (large interface relative to implementation, delegates). validateOrder should be identified as deep (small interface, focused implementation).",
"assertions": [
"Identifies processOrder as shallow using depth principles",
"Identifies validateOrder as deep using depth principles",
"Uses correct glossary terms (module, interface, depth)",
"Does not use terms 'boundary', 'component', or 'service' to describe depth"
]
},
{
"name": "domain-modeling-context-update",
"description": "Verify domain-modeling correctly updates CONTEXT.md with resolved terms",
"test_input": "During a conversation about an e-commerce system, the user says 'We should handle cancellation requests' but CONTEXT.md defines 'cancellation' as 'A request to void an order before shipment.' The user then clarifies 'I mean refunds — cancellation after delivery.'",
"expected_output": "The agent should: 1) Call out the conflict immediately. 2) Propose a new term 'refund' for post-delivery returns. 3) Update CONTEXT.md with the new term. 4) Not create an ADR for this (easy to reverse decision).",
"assertions": [
"Calls out the conflict between user's usage and glossary definition",
"Proposes 'refund' or similar precise canonical term for post-delivery returns",
"Updates CONTEXT.md inline (not batched)",
"Does NOT offer to create an ADR for this change"
]
},
{
"name": "code-review-two-axis-separation",
"description": "Verify code-review keeps Standards and Spec axes separate",
"test_input": "A PR adds a new API endpoint using a caching library the repo has never used before. The code compiles and passes tests, and implements the spec accurately. However, it uses a cache library not documented in any CODING_STANDARDS file. The spec says 'Implement caching for the product endpoint.'",
"expected_output": "Two-axis output: Standards axis flags the undocumented library choice; Spec axis confirms the feature matches requirements. Findings are NOT merged or reranked.",
"assertions": [
"Identifies the undocumented caching choice under Standards (not Spec)",
"Confirms the spec requirements are met under Spec (not Standards)",
"Presents findings under separate '## Standards' and '## Spec' headings",
"Does NOT merge or rerank findings across axes",
"Includes one-line summary with worst issue per axis (not a single winner)"
]
},
{
"name": "diagnosing-bugs-feedback-loop-first",
"description": "Verify diagnosing-bugs builds feedback loop before hypothesizing",
"test_input": "User reports: 'The checkout page throws a 500 error when applying a coupon code. I don't have a test for this. Here's the error in production logs.'",
"expected_output": "Phase 1 delivers a tight, red-capable loop BEFORE any hypothesis or code-reading theory. The loop goes red on the coupon 500 error specifically (not just 'page broke').",
"assertions": [
"Does NOT jump to hypothesis before building a feedback loop",
"Constructs a feedback loop that is red-capable (asserts the coupon 500 error)",
"Attempts to tighten the loop (faster, sharper signal, deterministic)",
"Redacts any secrets in captured output",
"Shows the loop being run and confirmed red before proceeding to Phase 2"
]
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation). | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK). | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |