Skip to content

pradeepmouli/skillit / skills/skillit-core

skillit-core

Verified

Shared types, SKILL.md renderer, and token budgeting for skillit plugins

@skillit/core

Shared types, SKILL.md renderer, and token budgeting for skillit plugins

When to Use

  • You have one or more ExtractedSkill objects and need SKILL.md + references/ output
  • Building a custom extraction pipeline that bypasses the TypeDoc plugin
  • You need fine-grained control over rendering a single skill
  • You have RenderedSkill objects from renderSkills() and need to persist them to the filesystem
  • Building a custom pipeline that separates rendering from writing (e.g., for preview or dry-run)
  • You need structured README sections for audit context or skill enrichment
  • The audit engine calls this to check for Features and Troubleshooting sections
  • You want programmatic quality feedback on JSDoc coverage before publishing skills
  • Building a CI gate that blocks PRs with undocumented exports
  • You have hand-written prose docs (tutorials, guides, architecture) in a docs/ directory
  • You want these docs included alongside API skills for richer agent context
  • You have ParsedMarkdownDoc objects from scanDocs() and need to attach them to an ExtractedSkill
  • You want to include well-known root-level markdown files (ARCHITECTURE.md, MIGRATION.md, CONTRIBUTING.md) as documents
  • Supplementing API skills with project-level documentation context
  • You have an examples/ directory alongside source code
  • You want to link example files to the exported symbols they demonstrate
  • You have scanned examples with scanExamples and want to attach them to
  • the relevant exported symbols inside an ExtractedSkill

Avoid when:

  • Rapid local iteration where audit noise slows you down — use the skillsAudit: false option instead
  • Your docs/ directory contains only auto-generated API docs — they duplicate the TypeDoc output
  • Your root directory has many auto-generated or release-tracking markdown files — use explicit include lists instead
  • Your examples are embedded as
  • API surface: 18 functions, 25 types

NEVER

  • NEVER set maxTokens below 500 — reference files become truncated mid-signature, producing broken code blocks that confuse LLMs
  • NEVER pass skills with empty name — the output directory becomes a bare / path
  • NEVER set maxTokens below 500 — reference files become truncated mid-signature, producing broken code blocks

Configuration

3 configuration interfaces — see references/config.md for details.

  • SkillRenderOptions — Options controlling skill rendering
  • LlmsTxtOptions
  • DocsExtractionOptions — Configuration options that control how a docs directory is scanned and which markdown files are included in the extraction.

Quick Reference

Rendering: renderSkills, renderSkill I/O: writeSkills Token Management: estimateTokens, truncateToTokenBudget llms-txt: renderLlmsTxt, LlmsTxtResult Parsing: parseReadme, parseMarkdownDoc, scanDocs, docsToExtractedDocuments, scanRootDocs, scanExamples, linkExamplesToSkill Audit: auditSkill, formatAuditText, formatAuditJson config-renderer: renderConfigSurfaceSection, renderConfigReference types: ExtractedSkill, ExtractedFunction, ExtractedClass, ExtractedType, ExtractedEnum, ExtractedParameter, ExtractedProperty, ExtractedVariable, ExtractedDocument, RenderedFile, RenderedSkill audit-types: AuditSeverity, AuditIssue, AuditPass, AuditContext, ParsedReadme, AuditResult config-types: ConfigSourceType, ExtractedConfigSurface, ExtractedConfigOption, ExtractedConfigArgument markdown-types: ParsedSection, ParsedMarkdownDoc examples-scanner: ParsedExample

Links

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: 469kUpdated Claude CodeCodexVerifiedInstall
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: 321kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 321kUpdated Claude CodeCodexVerifiedInstall
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 availableStars: 321kUpdated Claude CodeCodexVerifiedInstall