Skip to content

hashgraph-online/awesome-codex-plugins / plugins/alpoxdev/hypercore/skills/hyper-skill

hyper-skill

Flagged

Use this skill when the user asks to create or refactor a reusable agent skill folder, including SKILL.md trigger wording, instruction contracts, rules, references, scripts, assets, and validation checks. Do not use for generic documentation that is not a skill.

rules/self-containment.md

View on GitHub
# Self-Containment

**Purpose**: Keep a skill usable from its own folder alone, so no other skill has to be installed, invoked, or present for the skill to work.

A skill is a distribution unit. If it points at another skill - by name, by path, or by requiring that skill's command to run - then a reader who installs only this skill gets a dangling instruction. The fix is never "install the neighbor too"; it is to bring what is needed into this folder or to describe the boundary without naming the neighbor.

## 1. Normative Rules

- SK-S-1: A skill is whole with its own folder. It must not require another skill's installation, invocation, path, or documents. When content from elsewhere is genuinely needed, copy it into this skill's own `references/` as a local snapshot instead of pointing at the other artifact. A repository-maintenance command that contains another skill's path is a representative case: it does not belong in the skill artifact at all, because repository policy documents own it.
- SK-S-2: Describe boundaries by output shape, not by neighbor name. "Use this skill when the output is a reusable skill folder; general documents, runbooks, prompts, and plans are out of scope" keeps the trigger boundary intact without naming another skill. Naming a neighbor creates a reference the reader cannot resolve.
- SK-S-3: A user instruction is the only exception. When the user explicitly asks for a cross-skill reference, run the check with `--allow <sibling-name>` for that name, keep the finding visible as a warning, and record the instruction in the task record. The exception never becomes the default.

## 2. Detection Rules

`skills/hyper-skill/scripts/validate-hyper-skill.sh` carries `checkSelfContainment`, which scans a package's text files for:

- a hyphenated sibling skill name as a bare or backticked token, with boundaries that exclude longer identifiers, so an extended name is not misread as its prefix;
- any sibling skill name in `$<name>` invocation form;
- any sibling skill name as a `skills/<name>/` path segment.

Sibling names come from the directories beside the package root; a directory counts only when it carries a `SKILL.md`. In the default mode the check reports errors for a package that adopts this contract and warnings for any other package. `--require-self-containment` promotes findings to errors regardless of package name, which is the mode to use on a skill this authoring skill has just produced, and it also widens the path rule: any `skills/<name>/` path other than the package's own is reported even when that skill is not installed beside this package, so a standalone or moved package is still caught. The `$<name>` form stays sibling-based in both modes, because vendor runtime built-ins use the same syntax. `--allow <sibling-name>` (repeatable) records a user-directed exception.

## 3. Limitations

- A single-word sibling name is not detectable as prose or in backticks, because it is indistinguishable from an ordinary word that also appears as a capability name. Only its `$<name>` and `skills/<name>/` forms are detected. Authors avoid the prose form by hand.
- Detection reads text, so it cannot see a reference that is assembled at run time from fragments.
- The check proves the absence of the shapes it knows. It does not prove that the skill works alone; that still needs the skill to be read and run on its own.

## Sources

> Links checked 2026-09-21.

| Claim | Source |
|---|---|
| The self-containment requirement, the output-shape boundary rule, and the user-instruction-only exception | user instruction, 2026-09-21 |
| The detection rules, the `--require-self-containment` and `--allow` flags, and the sibling-name discovery | `skills/hyper-skill/scripts/validate-hyper-skill.sh` |
| One canonical home per rule and a lean core, which is why this rule lives in one file instead of being restated in the core | `rules/progressive-disclosure.md` |

### Evidence grade

`LOCAL` - this rule states repository policy and a user instruction. It is not an external claim, and no vendor source is cited for it.
rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization.Stars: 469kUpdated Claude CodeCodexVerifiedInstall
academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.Stars: 197kUpdated Claude CodeCodexVerifiedInstall
claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file).Stars: 197kUpdated Claude CodeCodexUnverifiedInstall
doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.Stars: 197kUpdated Claude CodeCodexVerifiedInstall