Conradgui/matt-pocock-inspired-skill-writing / skills/skill-writing-guidelines
skill-writing-guidelines
VerifiedWrite or review agent skills for predictable invocation and execution. Use when creating a skill, restructuring a skill collection, diagnosing unreliable skill behavior, or pruning an existing skill.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.8 KB |
| agents/openai.yaml | config | 252 B |
| references/evidence-boundary.md | markdown | 1.8 KB |
references/evidence-boundary.md
View on GitHub# Evidence boundary
This skill is an independent interpretation inspired by the public
[`mattpocock/skills`](https://github.com/mattpocock/skills) repository. It is not created,
maintained, or endorsed by Matt Pocock.
The study used
[`mattpocock/skills@66898f60e8c744e269f8ce06c2b2b99ce7660d5f`](https://github.com/mattpocock/skills/tree/66898f60e8c744e269f8ce06c2b2b99ce7660d5f),
covering all 39 `SKILL.md` files plus directly linked references, repository routing, ADRs, and
changelog evidence.
Use these attribution categories:
- **Direct principle**: explicitly documented by the source repository.
- **Repository pattern**: repeatedly demonstrated across independent source skills.
- **Interpretation**: a conclusion synthesized from those patterns.
The twelve-part organization is an interpretation. Do not describe it as Matt Pocock's official
list. Harness fields are also non-portable: translate invocation and packaging policy to the active
platform instead of copying source metadata unchanged.
“Proportional rigor,” the smallest-decision-changing evidence rule, and its stopping condition are
this project's synthesis from source cost tradeoffs, fail-fast and early-exit patterns, hard/soft
dependency policy, YAGNI scoping, pruning, and fact-versus-decision boundaries. They are not a named
Matt Pocock framework or a thirteenth source principle.
The source snapshot demonstrates lifecycle buckets, per-skill Codex metadata, a curated plugin
manifest, setup/provider adapters, human documentation policy, Changesets, and release automation.
It does not contain Python files, a Python compilation gate, a repository-wide skill validator, a
universal lint/test suite, or a stated numeric `SKILL.md` limit. Treat those stronger release gates
as independent extensions when a project's risk warrants them.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |