matt-mode
Use when the user invokes matt-mode to clarify requirements, discuss design, map decisions, or prepare implementation through specs and tickets.
- License 许可
- MIT
- 最近更新
- 2026-11-15
- Stars
- ★ 0
- Forks
- 1
01 前言 Front Matter
- kind
- dev
- name
- matt-mode
- description
- Use when the user invokes matt-mode to clarify requirements, discuss design, map decisions, or prepare implementation through specs and tickets.
02 资源文件 25
- 参考文档agents/openai.yaml209 B
- 参考文档playbooks/codebase-design/codebase-design.md6.3 KB
- 参考文档playbooks/codebase-design/DEEPENING.md2.5 KB
- 参考文档playbooks/codebase-design/DESIGN-IT-TWICE.md2.6 KB
- 参考文档playbooks/domain-modeling/ADR-FORMAT.md2.7 KB
- 参考文档playbooks/domain-modeling/CONTEXT-FORMAT.md2.2 KB
- 参考文档playbooks/domain-modeling/domain-modeling.md3.2 KB
- 参考文档playbooks/grill-me/grill-me.md111 B
- 参考文档playbooks/grill-with-docs/grill-with-docs.md194 B
- 参考文档playbooks/to-spec/to-spec.md2.9 KB
- 参考文档playbooks/to-tickets/to-tickets.md5.5 KB
- 参考文档playbooks/wayfinder/wayfinder.md11.6 KB
- 参考文档references/acceptance-cases.md5.1 KB
- 许可证references/LICENSE1.0 KB
- 参考文档references/lineage.md6.6 KB
- 参考文档references/local-adaptations.md5.9 KB
- 参考文档references/pstack-handoff.md738 B
- 参考文档references/workspace.md4.4 KB
- 脚本scripts/check_matt_mode.py19.8 KB
- 脚本scripts/check_meta_skill_layout.py41.2 KB
- 脚本scripts/tests/test_check_matt_mode.py1.3 KB
- 脚本scripts/tests/test_check_meta_skill_layout.py2.3 KB
- 脚本scripts/tests/test_update_matt_mode.py7.8 KB
- 脚本scripts/update_matt_mode.py40.8 KB
- 参考文档upstream-lock.json7.0 KB
03 SKILL.md 正文
Matt mode
Matt-mode prepares implementation. It produces planning artifacts and implementation-ready context; it does not implement software, including through subagents.
Use Matt Pocock's procedures under their original names. Read local adaptations, then the procedure for the requested result. Resolve bundled paths relative to this file. Load supporting files only when the selected procedure needs them.
The procedure bodies and their supporting files are imported from a pinned upstream revision. Follow their full instructions within the local adaptations and the caller's scope. Keep routing follow-ups for the same effort until the user changes task or mode. Stop at the requested result.
Choose a procedure
Recognize spaced names such as "to spec", "to tickets", "grill with docs", and "domain modeling" as their hyphenated route names. A caller can resolve this package and read one named procedure without changing its own workflow or activating every route.
| Name | Use when | Read |
|---|---|---|
grill-me |
Sharpen an idea through conversation | grill-me |
grill-with-docs |
Interview while recording resolved domain terms and qualifying ADRs | grill-with-docs |
domain-modeling |
Define or challenge domain vocabulary and relationships | domain-modeling |
codebase-design |
Design a module's interface, depth, or seam | codebase-design |
wayfinder |
Chart or resume decisions across sessions while the route is uncertain | wayfinder |
to-spec |
Synthesize existing context into a spec, including design and testing decisions | to-spec |
to-tickets |
Break defined work into verifiable deliveries and blocking edges | to-tickets |
For an architecture review of existing code, including improve-codebase-architecture, direct the caller to code-review-mode instead of running a Matt-mode procedure.
When upstream asks to call a Skill by one of these names, read that procedure here. A slash command is an instruction pointer, not a required executable or separate installed skill.
Shared skills
Resolve research, grilling, and writing-for-agents from the active catalog or the sibling skill directories in a source or compiled installation. Read each skill's SKILL.md before using it. Their canonical upstream bodies live in those packages, not here.
researchgathers primary-source evidence, including when the user saysmatt-mode ; researchgrillingsupplies the interview used bygrill-me,grill-with-docs, and Wayfinderwriting-for-agentssupplies the writing discipline when producing a spec, tickets, a map, or another artifact an agent will consume
If a required dependency is unavailable, name it and the affected step. Continue independent work without claiming the missing procedure ran.
Choose the order
Small, obvious changes can go directly to the execution workflow. Use grill-me for discussion, or grill-with-docs when documentation is requested. Domain modeling and codebase design apply within relevant work, not as compulsory preliminary stages.
Research supplies missing facts through source inspection and primary documentation. Wayfinder manages uncertainty that exceeds one session; defined work does not need a map just because it is large.
For a substantial planning effort, the usual sequence is discussion, then to-spec, then to-tickets. Skip completed stages. A to-spec request synthesizes the available context without restarting the interview.
Finish with the requested spec, plan, tickets, or readiness assessment. When implementation or a runnable prototype is requested, direct the user to poteto-mode with the existing context. Read the PStack handoff when preparing that context; completion of planning does not start execution.
Maintenance
Read lineage and updating to refresh upstream content. Use acceptance cases when changing routing or local adaptations. Neither is needed during ordinary work.
04 同仓库更多 Skills
writing-great-skills
Use when creating, modifying, evaluating a skill.
writing-for-agents
Use when creating, editing, or reviewing a skill, AGENTS.md, CLAUDE.md, or another document agents read.
why
Use for questions about design rationale, the history of regressions or incidents, or the evidence behind thresholds and tradeoffs. Use `how` for runtime behavior. Do not use for plain commit or date lookups.
verify-video-archive
Use when validating the macOS or Linux archive workflow reached by `just convert-video`, including real media, prerequisites, terminal progress, source safety, locking, and recovery.
verify-transcript
Use when validating transcript CLI behavior, locating its verification features, or running the paid YouTube end-to-end check.
verify-skills
Use when verifying `just compile-skills`, `just remote-skills`, `just install-skills`, or `just skills-discover` in the skills repository, such as after changing their scripts or adding, renaming, or moving a skill.