# AI use cases for Skills (mattpocock/skills) Repo: [mattpocock/skills](https://github.com/mattpocock/skills) | Use cases: 5 | Page: https://dedupedai.com/use-cases/mattpocock-skills/ A set of editable agent skills for planning, aligning, building, and debugging software with coding agents, taken from the author's .agents directory. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Align an agent on a change before any code is written: https://dedupedai.com/use-cases/mattpocock-skills/align-agent-on-change-before-coding/ - Split a plan into tracer-bullet tickets with blockers: https://dedupedai.com/use-cases/mattpocock-skills/split-a-plan-into-blocking-tickets/ - Fix a bug with a failing test before the fix: https://dedupedai.com/use-cases/mattpocock-skills/fix-a-bug-with-a-failing-test-first/ - Find deepening opportunities in an aging codebase: https://dedupedai.com/use-cases/mattpocock-skills/survey-codebase-for-deepening-opportunities/ - Plan a huge project as a map of decision tickets: https://dedupedai.com/use-cases/mattpocock-skills/plan-a-huge-project-as-decision-tickets/ ## How to align an agent on a change before any code is written with Skills Repo: [mattpocock/skills](https://github.com/mattpocock/skills) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Run a grilling session so the agent asks detailed questions before it edits your repo. **Who it is for:** A developer about to ask a coding agent for a feature who has seen past builds miss what they meant. **The problem:** Misalignment is the most common failure mode: the agent builds something that does not match what you wanted. **The outcome:** A clarified plan, plus shared terms and decisions recorded in the project's domain docs before implementation starts. ### Steps 1. Install the plugin in Claude Code with `claude plugin install mattpocock-skills@claude-plugins-official`. 2. In your repo, run `/setup-matt-pocock-skills` once and answer the questions about issue tracker, triage labels, and where docs go. 3. Run `/grill-with-docs` and describe the change you want. 4. Answer the agent's questions until the plan is clear, and let it update the glossary and ADRs as you go. ### Prompt to paste into your agent ```text In the mattpocock/skills repo's grill-with-docs workflow, interview me about the change described here: . Expected output: a short list of resolved decisions and updated GLOSSARY.md entries or ADRs. Verify that every new term in the output appears in the glossary file. ``` ### Caveat The grilling session only helps if you engage with the questions, and the Skills expect the setup step to run once per repo before the engineering Skills are used. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/mattpocock-skills/align-agent-on-change-before-coding/ ## How to split a plan into tracer-bullet tickets with blockers with Skills Repo: [mattpocock/skills](https://github.com/mattpocock/skills) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Turn a conversation into a spec, then into ordered tickets that declare what blocks what. **Who it is for:** A tech lead or solo developer with a big feature idea who needs work broken into pieces an agent can pick up. **The problem:** Large changes stall when the work is not sliced, and agents lose track of what must finish first. **The outcome:** A published spec and a set of tracer-bullet tickets, each with its blocking edges, in your tracker or a local file. ### Steps 1. Install the plugin in Codex with `codex plugin marketplace add mattpocock/skills`, then run `codex plugin add mattpocock-skills@mattpocock`. 2. Run `/setup-matt-pocock-skills` once in the repo and choose your issue tracker or local files. 3. Discuss the feature with the agent, then run `/to-spec` to publish the spec from the conversation. 4. Run `/to-tickets` on the spec and pick either local text files or native blocking links on your tracker. ### Prompt to paste into your agent ```text In the mattpocock/skills repo, use to-spec on our current discussion about , then use to-tickets to split the spec into tracer-bullet tickets. Expected output: a published spec and a ticket list where each ticket names its blocking tickets. Verify that no ticket is blocked by a ticket that depends on it. ``` ### Caveat Native blocking links are only available when you use a real tracker; with local files the blocking edges are written as text. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/mattpocock-skills/split-a-plan-into-blocking-tickets/ ## How to fix a bug with a failing test before the fix with Skills Repo: [mattpocock/skills](https://github.com/mattpocock/skills) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Diagnose a bug in phases, then drive a red-green-refactor loop to fix it. **Who it is for:** A developer whose agent keeps producing code that does not work and who needs a disciplined debugging loop. **The problem:** Without feedback on how code actually runs, the agent guesses, and fixes often miss the real cause. **The outcome:** A written reproduction, a test that fails before the fix and passes after, and a refactored fix. ### Steps 1. Install the plugin in Claude Code with `claude plugin install mattpocock-skills@claude-plugins-official`. 2. Describe the failing behavior and run `/diagnosing-bugs`, working through each gated phase. 3. Run `/tdd` to write a failing test for the bug first. 4. Make the test pass, then refactor while keeping it green. ### Prompt to paste into your agent ```text In the mattpocock/skills repo, use diagnosing-bugs on this failure: , then use tdd to write a failing test before any fix. Expected output: a root-cause note and a red test that turns green after the fix. Verify that the test fails on the unfixed code before you accept the change. ``` ### Caveat The red-green-refactor loop assumes your project already has a test setup the agent can run. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/mattpocock-skills/fix-a-bug-with-a-failing-test-first/ ## How to find deepening opportunities in an aging codebase with Skills Repo: [mattpocock/skills](https://github.com/mattpocock/skills) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Scan a repo for modules to deepen, review an HTML report, then grill through one pick. **Who it is for:** A developer maintaining a tangled codebase that agents have been speeding up into more complexity. **The problem:** Agent-driven coding accelerates software entropy, so the design drifts faster than anyone reviews it. **The outcome:** A visual HTML report of candidate deepening opportunities and a grilled plan for the one you choose. ### Steps 1. Install the plugin in Claude Code with `claude plugin install mattpocock-skills@claude-plugins-official`. 2. Run `/improve-codebase-architecture` on your repo. 3. Open the HTML report and pick one candidate. 4. Let the agent grill you through the chosen candidate before it changes any code. ### Prompt to paste into your agent ```text In the mattpocock/skills repo, run improve-codebase-architecture on . Expected output: an HTML report of deepening candidates, then a grilling session on the one I pick. Verify that the report lists concrete modules rather than general advice. ``` ### Caveat This is a survey, not a rescue: on a genuinely old codebase it finds real candidates but will not untangle the mud for you. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/mattpocock-skills/survey-codebase-for-deepening-opportunities/ ## How to plan a huge project as a map of decision tickets with Skills Repo: [mattpocock/skills](https://github.com/mattpocock/skills) | Difficulty: Advanced | Time: 60 minutes | Basis: From the README Use wayfinder to resolve decisions one at a time across several agent sessions. **Who it is for:** A lead planning a project too large for one agent session who needs a shared map of open decisions. **The problem:** Work bigger than one session loses context, and open decisions block progress without anyone tracking them. **The outcome:** A shared map of decision tickets on the issue tracker, with each decision resolved until the route to the destination is clear. ### Steps 1. Install the plugin in Codex with `codex plugin marketplace add mattpocock/skills`, then run `codex plugin add mattpocock-skills@mattpocock`. 2. Run `/setup-matt-pocock-skills` once in the repo and connect your issue tracker. 3. Run `/wayfinder` and state the destination you want to reach. 4. Resolve the decision tickets one at a time in later sessions until the way forward is clear. ### Prompt to paste into your agent ```text In the mattpocock/skills repo, use wayfinder to map the work needed to reach . Expected output: a set of decision tickets on the issue tracker and the first one to resolve. Verify that each decision ticket is stored in the tracker and not only in the chat. ``` ### Caveat Wayfinder keeps its map on the issue tracker, so it needs a real tracker configured during setup. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/mattpocock-skills/plan-a-huge-project-as-decision-tickets/