All skills
simota avatar

/sigil

@35ffd55
by shingo imotasimota/agent-skills85 stars
15

Designing a repository's project-local operating layer and generating its skills, recipes, workflows, and routing map. Not for global ecosystem agents (Architect) or runtime execution (Nexus).

Use this Skill: https://skilld.dev/gh/simota/agent-skills/sigil

This session only. Nothing lands on disk.

referencecontext-analysis.md

≈2k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Context Analysis

Purpose: load this during SCAN to detect the project's real stack, conventions, rule files, existing skills, and evolution signals before generating anything.

Contents

  1. Tech stack detection
  2. Structure and naming
  3. Monorepo and polyglot handling
  4. Rules and config precedence
  5. Existing skill audit
  6. Style-profile extraction

Tech Stack Detection

Manifest File Mapping

File Stack Inspect for
package.json Node.js / JavaScript / TypeScript dependencies, devDependencies, scripts, workspaces
tsconfig.json TypeScript strictness, paths, baseUrl
go.mod Go module path, dependencies
Cargo.toml Rust dependencies, features
pyproject.toml / setup.py / requirements.txt Python framework and tool dependencies
Gemfile Ruby gems and Rails presence
pom.xml / build.gradle Java / Kotlin starters, plugins
composer.json PHP framework and autoloading
pubspec.yaml Dart / Flutter packages and SDK
mix.exs Elixir deps
bun.lockb / bunfig.toml Bun Bun runtime usage

Framework Detection

Framework Detection signal Likely priority skills
Next.js next dependency + next.config.* new-page, new-api-route, new-component
React (Vite / CRA) react without next new-component, new-hook, new-context
Vue.js vue dependency new-component, new-composable, new-store
Nuxt nuxt dependency new-page, new-server-route, new-composable
Remix @remix-run/node dependency new-route, new-loader, new-action
Express express dependency new-route, new-middleware, new-controller
Fastify fastify dependency new-route, new-plugin, new-schema
Hono hono dependency new-route, new-middleware, new-validator
NestJS @nestjs/core dependency new-module, new-controller, new-service
tRPC @trpc/server dependency new-router, new-procedure, new-middleware
FastAPI fastapi in Python manifests new-router, new-model, new-schema
Django django in Python manifests new-app, new-model, new-view
Rails rails in Gemfile new-model, new-controller, new-migration
Go stdlib / Chi / Echo go.mod without larger framework or with router deps new-handler, new-middleware, new-model
Gin github.com/gin-gonic/gin new-handler, new-middleware
Actix / Axum Rust web framework in Cargo.toml new-handler, new-service
Spring Boot spring-boot-starter* new-controller, new-service, new-repository
Laravel laravel/framework new-controller, new-model, new-migration
Flutter flutter SDK new-screen, new-widget, new-bloc
SvelteKit @sveltejs/kit new-route, new-component, new-server

Structure and Naming

Directory Signals

Pattern Meaning Skill impact
src/app/ App Router or route-first structure Route-oriented templates
src/pages/ Pages Router or Nuxt-style routing File-based route templates
src/components/ Shared components Component creation skills
src/hooks/ / src/composables/ Reusable logic layer Hook/composable skills
src/lib/ / src/utils/ Utility helpers Utility templates
src/services/ Service layer present Service and integration skills
src/stores/ / src/store/ Centralized state State-management skills
__tests__/ / *.test.* / *.spec.* Test layout Test placement and naming
e2e/ / cypress/ / playwright/ E2E test framework End-to-end testing conventions
prisma/ / drizzle/ / migrations/ Database stack Schema and migration skills
cmd/ / internal/ Go package layout CLI and internal package patterns

Naming Detection

Check at least 3+ existing files of the same kind before templating:

  1. File names: kebab-case, camelCase, PascalCase, snake_case
  2. Component names: Button.tsx vs button/index.tsx
  3. Test layout: colocated vs separate __tests__/
  4. Import style: relative imports, aliases, barrel exports

Monorepo and Polyglot Handling

Monorepo Detection

File Tool Scope signal
turbo.json Turborepo pipeline and package graph
nx.json Nx projects and target defaults
pnpm-workspace.yaml pnpm workspaces workspace globs
lerna.json Lerna managed packages
root package.json with workspaces npm/yarn workspaces workspace entries

Monorepo Strategy

  1. Scan the root to detect the monorepo tool and package graph.
  2. Run full SCAN per package.
  3. Generate shared root skills for cross-cutting concerns only.
  4. Generate package-specific skills for framework-specific workflows.
  5. Never duplicate a root skill inside a package.

Package Priority Order

  1. Primary application packages
  2. API / backend packages
  3. Shared libraries
  4. Tooling packages

Polyglot Detection

Pattern Strategy
package.json + go.mod Separate JS/TS and Go skill sets; share conventions
package.json + pyproject.toml Separate JS/TS and Python skill sets; share conventions
Cargo.toml + package.json Add bridge skills only for interop boundaries
Multiple manifests in subdirectories Scope skills per directory or package

For ambiguous names, prefix the skill directory with the language: ts-new-component, go-new-handler, py-new-router.

Rules and Config Precedence

Convention Priority

  1. Explicit config and rule files
  2. Existing code patterns
  3. Framework defaults
  4. Community standards

Rule Files to Inspect

  • CLAUDE.md
  • .cursorrules
  • .windsurfrules
  • AGENTS.md
  • .github/copilot-instructions.md

Config Files to Inspect

File group What to infer
.eslintrc*, eslint.config.*, .prettierrc*, .editorconfig naming, formatting, import expectations
tsconfig.json aliases, module resolution, strictness
Makefile, Taskfile.yml recurring commands worth skilling
docker-compose.yml service topology and local dev flow
.github/workflows/*.yml CI/CD and deployment workflows
.husky/, .lefthook.yml git-hook expectations

Existing Skills Audit

Before generating or updating skills, inspect:

  1. Project .claude/skills/
  2. Project .agents/skills/
  3. Project CLAUDE.md
  4. Ecosystem agents, to avoid overlapping core functions

Deduplication Rules

A skill already exists if any of these are true:

  • Same name exists in either project skill directory
  • Functional overlap is greater than 70%
  • It duplicates an ecosystem agent's core function

Directory Sync Check

Both directories must stay identical.

  1. List skills in .claude/skills/ and .agents/skills/
  2. Find orphans that exist in only one directory
  3. Copy the orphan to the missing directory before new generation
  4. Report every sync repair

Style-Profile Extraction

When the project already contains skills, learn their style before authoring new ones.

What to Sample

  • Section ordering and naming
  • Description tone
  • Heading language
  • Template depth
  • Example style
  • Checklist usage

Extraction Rules

  1. Read all existing skills in both directories.
  2. Use majority patterns when at least 10+ comparable files exist.
  3. Match language, structure, and depth unless the existing pattern is clearly low quality.
  4. If existing skills are low quality, note the drift and do not propagate the defect.
  5. If no existing skills exist, use Sigil defaults from skill-templates.md.

Source: SKILL.md on GitHub

1 warning13d5 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    The skill is a project-local development tool designed to analyze codebase context and generate AI agent skills. It uses shell commands and local file access to detect tech stacks and conventions. While it includes safety guardrails against prompt injection and credential exposure, it possesses an attack surface for indirect prompt injection, where a malicious repository could influence the generation of executable skill instructions or configurations.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    2/12 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 35ffd55. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 3 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/sigil