Skilllibrary skill-adaptation
Adapt a general-purpose skill to a specific repository, team, stack, or project context without losing the base pattern. Use this when porting a skill between stacks (e.g. React to Vue, npm to pnpm), customizing a library skill for project conventions, or localizing tool references for a different environment. Do not use for creating skills from scratch (use skill-authoring) or for minor refinements to an already project-specific skill (use skill-refinement).
install
source · Clone the upstream repo
git clone https://github.com/merceralex397-collab/skilllibrary
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/merceralex397-collab/skilllibrary "$T" && mkdir -p ~/.claude/skills && cp -r "$T/03-meta-skill-engineering/skill-adaptation" ~/.claude/skills/merceralex397-collab-skilllibrary-skill-adaptation && rm -rf "$T"
manifest:
03-meta-skill-engineering/skill-adaptation/SKILL.mdsource content
Purpose
Takes a general-purpose skill and adapts it for a specific context: different tech stack, different organization, different domain, or different conventions. Preserves the core procedure while replacing generic references with context-specific ones.
When to use this skill
Use when:
- User says "adapt this skill for our project", "customize for Python/React/etc.", "make this work for our team"
- Installing library skill that needs project-specific conventions
- Porting skill from one stack to another (e.g., React → Vue, npm → pnpm)
- Skill references tools/APIs/conventions that don't exist in target environment
Do NOT use when:
- Skill is already project-specific and just needs refinement (use
)skill-refinement - Creating skill from scratch (use
)skill-authoring - Only trigger description needs work (use
)skill-description-optimizer - Skill works fine in current context
Operating procedure
- Read the source skill completely: Understand the problem it solves, output it produces, and assumptions it encodes
- Identify the target context:
- Stack/language: What technologies?
- Conventions: Naming, file structure, coding style?
- Tools: What's available? What's forbidden?
- Domain: What terminology?
- Catalog adaptation points (things that MUST change):
- Tool/command references (e.g.,
→npm
,pnpm
→pytest
)unittest - File paths and patterns (e.g.,
→src/components
)app/ui - Naming conventions (e.g., camelCase → snake_case)
- Output formats (e.g., Markdown → org-mode)
- Domain terminology (e.g., "user" → "customer")
- Tool/command references (e.g.,
- Identify invariants (things that must NOT change):
- Core procedure logic
- Safety constraints
- Quality checks
- The fundamental problem being solved
- Create adapted version:
- Keep frontmatter structure
- Update description if target context affects routing
- Replace all adaptation points with target-specific versions
- Add context-specific examples if skill uses few-shot
- Validate the adaptation:
- Every tool reference exists in target context
- Every file path pattern makes sense
- No dangling references to source-context things
- Document what was adapted: Add Adaptation Note at bottom noting source skill and what changed
Output defaults
## Adaptation Summary **Source skill**: [name and location] **Target context**: [stack/project/team] ### Changes Made | Original | Adapted | Reason | |----------|---------|--------| | npm install | pnpm add | Project uses pnpm | | src/components/ | app/ui/ | Different structure | ### Invariants Preserved - [Core logic 1] - [Core logic 2] ### Adapted SKILL.md [Full content] ### Adaptation Note Adapted from [source] for [target]. Changes: [list].
References
- Original skill documentation
- Target project's conventions (AGENTS.md, README, etc.)
- https://docs.github.com/en/copilot/concepts/agents/about-agent-skills
- https://docs.github.com/en/copilot/how-tos/use-copilot-agents/coding-agent/create-skills
- https://docs.anthropic.com/en/docs/claude-code/overview
- https://developers.openai.com/codex/skills
Failure handling
- Source skill poorly documented: Infer intent from procedure, flag assumptions made
- Target context unclear: Ask: "What stack? What file structure? What tools available?"
- Adaptation breaks core logic: Skill may not be adaptable—recommend creating new skill
- Missing equivalent in target: Flag gap, suggest alternatives or note limitation