agentsclimarketplace

Skill injector

Skill d3x293/code-crew/skills/skill-injector

Multi-agent orchestration plugin for Claude Code — 11 AI agents with intelligent model routing, 3-layer codebase indexing, and auto skill injection

Install
npx -y skills add d3x293/code-crew --skill skill-injector

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 13 stars13 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Auto-selects and injects relevant skills from the 125+ ECC catalog into agent prompts based on task type, project stack, and agent role. Prevents skill overload by loading only what's needed.

SKILL.md

4.9 KB, as published. Nobody here has run it

Skill Injector - Smart Skill Selection

Maps task signals to skills from the everything-claude-code catalog and injects only the relevant ones into each agent's prompt. Prevents context bloat by being selective.

When to Activate

  • During task routing (called by task-router skill)
  • When an agent needs additional skill guidance mid-task

Skill Selection Algorithm

Input

  • Task classification (type + complexity) from task-router
  • Agent role being dispatched
  • Project stack from crew-profile.md
  • Skill catalog from crew-skills.json

Selection Rules

Rule 1: Always-Active Skills (injected into every agent) These are universal best practices (kept minimal for token efficiency):

  • context-budget — monitor token consumption (subsumes strategic-compact)
  • verification-loop — verify before claiming done
  • coding-standards — follow project conventions

Note: iterative-retrieval moved to bug-fix task-type skills. search-first moved to feature task-type skills. strategic-compact dropped (redundant with context-budget).

Rule 2: Stack-Specific Skills (based on detected languages)

StackSkills
JavaScript/TypeScripttypescript-patterns, backend-patterns
Reactreact-patterns, react-testing
Next.jsnextjs-patterns
Vuevue-patterns, nuxt4-patterns
Pythonpython-patterns
Djangodjango-patterns, django-security, django-tdd
Flaskpython-patterns
Gogolang-patterns
Rustrust-patterns
Javajava-patterns, springboot-patterns
Kotlinkotlin-patterns, kotlin-testing
C++cpp-coding-standards, cpp-testing
C#csharp-testing, dotnet-patterns
Flutter/Dartdart-flutter-patterns
Swiftswiftui-patterns, swift-concurrency-6-2
PHP/Laravelphp-patterns, laravel-patterns

Rule 3: Task-Type Skills (based on classified task type)

Task TypeSkills
bug-fixiterative-retrieval, deep-research, codebase-onboarding
featurearchitecture-decision-records, agentic-engineering, blueprint, search-first
refactorarchitecture-decision-records, backend-patterns
testtdd-workflow, e2e-testing, ai-regression-testing
securitysecurity-scanning
docsarticle-writing, documentation-lookup
devopsdeployment-patterns, docker-patterns, database-migrations, git-workflow
reviewverification-loop
performancecost-aware-llm-pipeline

Rule 4: Agent-Role Skills (specific to the dispatched agent)

AgentAdditional Skills
ceoagentic-engineering, architecture-decision-records
vp-engineeringarchitecture-decision-records, hexagonal-architecture
vp-qualityverification-loop, tdd-workflow
senior-dev(stack skills + task skills)
junior-devcoding-standards only (keep prompt small for Haiku)
code-reviewerverification-loop, security-scanning
debuggeriterative-retrieval, deep-research
security-analystsecurity-scanning
devops-engineerdeployment-patterns, docker-patterns
doc-writerarticle-writing
test-engineertdd-workflow, e2e-testing

Output: Skill Injection Payload

For each agent being dispatched, generate a skills section:

## Injected Skills for this Task

### From: {skill-name}
{Condensed key guidance from the skill - NOT the full SKILL.md, just the most relevant rules and patterns for this specific subtask}

### From: {skill-name-2}
{Condensed guidance}

Token Budget per Agent

To prevent context bloat from too many skills:

Agent ModelMax Skill TokensMax Skills
Opus (CEO)3000 tokens6 skills
Sonnet (leads/senior)2000 tokens4 skills
Sonnet (planner in lite mode)2500 tokens5 skills
Haiku (junior/devops/docs)800 tokens2 skills

Lite mode note: When --lite flag is used, no Opus agents are dispatched. The vp-engineering agent takes over planning duties and gets an expanded budget (2500 tokens, 5 skills) to compensate for the extra architectural context it needs.

If selected skills exceed the budget:

  1. Prioritize: always-active > task-type > stack > agent-role
  2. Truncate to key rules only (skip examples, anti-patterns)
  3. Drop lowest-priority skills

Skill Content Condensation

When injecting a skill, don't paste the entire SKILL.md. Extract only:

  1. The core rules/principles (2-5 bullet points)
  2. The most relevant pattern for this specific task
  3. Any critical anti-patterns to avoid

Example condensation:

Full SKILL.md: 200 lines, ~2000 tokens
Condensed injection: 15 lines, ~150 tokens

This keeps agent prompts focused and within token budget.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.