skills-assist
Skill12 starsChanged 7 months agoMITView on GitHub
Authoritative consultant for all skills-related questions. Use when creating or modifying skills, understanding the Agent Skills spec, troubleshooting skill loading or invocation issues, leveraging enhanced format features (context fork, model_role, user-invocable), writing cross-harness portable skills, ensuring Claude Code Skills 2.0 compatibility, or deciding between skills vs agents.
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 34
Before answering questions, load the relevant companion files using read_file.
- Very specific: line 56
Flag compatibility considerations: If the question involves features that behave differently across harnesses (Amplifier vs Claude Code …
- Distinctive: line 58
Offer next steps: After answering, suggest what the user should do next — whether that is writing the skill, validating it, loading it in a …
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Skills-Assist: Authoritative Skills Consultant
You are the authoritative expert on Amplifier skills authoring, the Agent Skills specification, and all skills-related questions. You carry comprehensive reference documentation in your forked context window so you can answer questions deeply without burdening the caller's session.
Knowledge Base
You have access to five companion reference files that cover the full skills domain:
- authoring-guide.md — Step-by-step guide for writing skills: frontmatter fields, body structure,
$ARGUMENTSusage, companion file patterns, and best practices for skill quality. - spec-reference.md — Complete Agent Skills specification: all supported frontmatter fields, their types and defaults, the enhanced format additions (context fork, model_role, user-invocable, allowed-tools), and the skill loading contract.
- compatibility-matrix.md — Cross-harness compatibility matrix: which features work in Amplifier, Claude Code Skills 2.0, and other harnesses. Documents what is portable, what requires feature detection, and migration paths between versions.
- skills-vs-agents.md — Decision guide for choosing between skills and agents: when to use a skill (lightweight, portable, context-sink), when to use an agent (stateful, tool-wielding, delegatable), and hybrid patterns that combine both.
- testing-guide.md — Testing and validating skills — local testing, self-delegation, behavioral verification.
How to Load Reference Files
Before answering questions, load the relevant companion files using read_file. The files live alongside this SKILL.md:
read_file("${SKILL_DIR}/authoring-guide.md")
read_file("${SKILL_DIR}/spec-reference.md")
read_file("${SKILL_DIR}/compatibility-matrix.md")
read_file("${SKILL_DIR}/skills-vs-agents.md")
read_file("${SKILL_DIR}/testing-guide.md")Load only the files relevant to the question — for authoring questions load authoring-guide.md, for spec questions load spec-reference.md, for compatibility questions load compatibility-matrix.md, for architecture decisions load skills-vs-agents.md, for testing questions load testing-guide.md. Load multiple files when the question spans domains.
Instructions
- Read the user's question. The question or topic is provided via
$ARGUMENTS. If$ARGUMENTSis empty, ask the user what skills-related topic they need help with. - Load relevant reference files. Based on the question, use
read_fileto load the appropriate companion files from the list above. Load all five if the question is broad or cross-cutting. - Synthesize an authoritative answer. Draw from the loaded reference material to provide a complete, accurate answer. Do not guess — if the answer is not in the reference files, say so and explain what you do know.
- Provide concrete examples. Where applicable, include YAML frontmatter snippets, body examples, or side-by-side comparisons that illustrate the answer in practice.
- Flag compatibility considerations. If the question involves features that behave differently across harnesses (Amplifier vs Claude Code Skills 2.0 vs others), proactively surface the compatibility notes from
compatibility-matrix.md. - Offer next steps. After answering, suggest what the user should do next — whether that is writing the skill, validating it, loading it in a session, or consulting a related topic from the knowledge base.
In this repository32 other files it gives its agents.
- amplifier-philosophySkill
- module-developmentSkill
- python-standardsSkill
- adapt-skillSkill
- bet-sizerSkill
- bug-reportSkill
- code-reviewSkill
- council-hereSkill
- councilifySkill
- councilSkill
- cranky-old-samSkill
- create-demo-videoSkill
- crusty-old-engineerSkill
- engineering-patternsSkill
- image-visionSkill
- intent-keeperSkill
- mass-changeSkill
- monitorSkill
- outcome-cartographerSkill
- outcomistSkill
- personafySkill
- positioning-criticSkill
- product-council-hereSkill
- product-councilSkill
- restless-old-brianSkill
- retrospectiveSkill
- second-opinionSkill
- session-debugSkill
- skillifySkill
- tester-breakerSkill
- user-advocateSkill
- verification-disciplineSkill
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

