cheez-search
Search code through the most precise semantic backend — prefer tilth MCP, with AST search (sg) for structural patterns. Use when the user asks to find a symbol, definition, caller, import, or text pattern — phrases like "where is X defined", "what calls Y", "find all usages of Z", "trace this function", "find the TODO comments", "search for this error string". Use even when the user says "grep", "rg", "ripgrep", "ag", "ack", "fd", or "find". Do NOT use for reading whole files (use cheez-read), editing code (use cheez-write), or running tests/builds.
- Installs
- 159
- Last 24h
- —
- Trending rank
- —
- Security
- Not audited
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/paulnsorensen/easy-cheese --skill cheez-search- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads cheez-search automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
SKILL.md
The instructions your agent reads when the skill is triggered.
cheez-search
Backend contract: use a semantic code-search backend, not blind text search. Choose the narrowest tool that answers the question.
Backend detection
Pick the backend by question type:
- LSP wins for type-grounded questions: definitions through imports, references with shadowing/re-exports, hover/type info, implementations, rename planning, and server-known code actions.
- AST search /
sgwins for syntax shapes: structural patterns with metavariables and repeated syntax-shaped codemods. - tilth MCP wins for broad source search: symbols, callers, imports, content, bounded regex, and dependency/blast-radius context in one fresh repo scan.
Note: every tilth tool takes cwd — your absolute checkout directory (the server's own cwd is frozen at startup). In Claude Code a hook injects cwd automatically; do not set it. On harnesses without the hook, pass cwd explicitly so relative paths and scopes resolve. There is no root parameter.
Examples
"Where is handleAuth defined?"
tilth_search(queries: [{query: "handleAuth"}], scope: "src/")
# Search: "handleAuth" in src/ — 6 matches (2 definitions, 4 usages)
## src/auth.ts:44-89 [definition]
→ [44-89] export fn handleAuth(req, res, next)
## src/routes/api.ts:34 [usage]
→ [34] router.use('/api/protected/*', handleAuth);
The [definition] tag answers the question; usages come along for free.
"What calls validateToken?"
tilth_search(queries: [{query: "validateToken", kind: "callers"}])
# Callers: "validateToken" — 3 call sites
## src/auth.ts:62 [usage] in handleAuth
→ [62] const claims = validateToken(token);
## src/middleware/admin.ts:18 [usage] in requireAdmin
→ [18] if (!validateToken(req.headers.authorization)) return next(403);
kind: "callers" filters out comments and strings — only real call sites.
"Find any TODO that mentions retries"
tilth_search(queries: [{query: "TODO.*retry", kind: "regex"}], scope: "src/")
Use kind: "regex" for pattern matches across content; bound the scope to
keep the cost down.
Scope: when to use the code-search backend, when not
cheez-search owns code in tracked, parseable source files (Rust, TypeScript/TSX/JS, Python, Go, Java, Scala, C/C++, Ruby, PHP, C#, Swift). Symbol-shaped queries, callers, content, bounded regex, AST-shape patterns, and type-grounded lookups stay in an AST/LSP backend.
Scope and freshness
Use the backend scoped to the active repository and fresh on disk. tilth MCP walks the working tree on demand and respects .gitignore; native sg and LSP qualify when they see the current checkout and answer the requested semantic shape.
Practical consequences:
- No stale indexes: if a backend needs a refresh before it sees current files, refresh or choose a fresher backend.
- Do not use repo-scoped backends for sibling worktrees, system paths, or dependency caches unless the backend explicitly supports that scope.
- Route cross-repo questions before entering cheez-search.
When NOT to invoke /cheez-search
These questions are out of scope -- don't enter cheez-search for them; the table tells the calling workflow skill where to route instead.
| Question (don't use cheez-search) | Route to | Why |
|---|---|---|
Pattern with metavars (JSON.parse(JSON.stringify($X))) | sg (ast-grep) — sanctioned escape, callable from cheez-search via Bash | AST shapes name/text search can't express |
External library docs ("how does React's useEffect work?") | /briesearch (Context7) | Not your code; live vendor docs |
| Plain non-code text at scale (logs, build outputs, large CSVs) | host Bash with rg, jq, awk, head/tail from the calling workflow skill | Tree-sitter parsing wastes tokens here; format-specific tools win |
Files outside the repo (system paths, ~/Library, /etc) | host Grep / Bash from the calling workflow skill | cheez-search is repo-scoped (see above) |
| Commit, push, ordinary PR, or stacked PR mutation | /plate | Publication is not code search |
Route out before entering (type-grounded)
Name-shaped queries stay in the semantic search backend chosen for this run (tilth, native AST search, or an equivalent indexed source-search path). For type-grounded questions, route out before entering cheez-search -- see references/routing.md for LSP and Serena routing tables.
Choose your search kind
For code navigation (where is X / what calls Y / blast radius): start with kind:symbol to find the definition, then kind:callers for call sites. Fall to content/regex only when you don't have a symbol name.
| Goal | Example backend | Example |
|---|---|---|
| Find where a symbol is defined / used | tilth_search symbol kind, native AST symbol search, or LSP definition when type-grounded | See Examples — "Where is handleAuth defined?" |
| Find every call site of a function | tilth_search callers kind, native AST caller search, or LSP references when type-grounded | See Examples — "What calls validateToken?" |
| Find literal strings, TODOs, error messages | content search in the semantic backend | tilth_search(queries: [{query: "error message", kind: "content"}]) |
| Find lines matching a regex | bounded regex search in the semantic backend | tilth_search(queries: [{query: "rate.?limit", kind: "regex"}]) |
| Match an AST shape (template with metavars) | sg (ast-grep, via Bash) | sg --lang typescript -p 'JSON.parse(JSON.stringify($X))' --json src/ |
| Module import / blast-radius graph | tilth_deps, LSP references/import graph, or equivalent dependency context | tilth_deps(path: "src/auth.ts") |
Rule of thumb: stay in a semantic source-search backend for name-shaped or text-shaped queries.
Drop to sg only when the pattern needs structural metavariables ($X,
$$$BODY) that the selected backend can't express.
MCP Tool Reference
Full invocation shapes, parameter options, and per-kind worked examples are in
references/tilth-search-reference.md.
AST-shape Patterns -- ast-grep fallback
For shapes with metavars ($X, $$$BODY) drop to sg via Bash -- the only sanctioned shell escape, covering structural codemods via sg --rewrite. For syntax, the language matrix, safe-invocation rules, pitfalls, and the codemod dry-run protocol, see references/sg-patterns.md.
DO NOT and Scope Limits
- Semantic backend for code search.
sgcovers AST-shape metavar patterns and codemods; LSP covers type-grounded definitions and references (renames and code actions are edits — route them to cheez-write);findstays fine for non-name predicates (size, mtime, perms) outside code work. Do not use grep, rg, ripgrep, ag, ack, find, or fd for code. - Expanded results are already shown. Do not re-read them.
- Start with the default expand count. Increase only if needed; do not overuse expand.
- Read entire files via cheez-read. Not here.
- Edit code via cheez-write. Not here.
- Run tests via the project's test/build skills. Not here.
- Handle git operations via git/gh skills. Not here.
FAQ
What does the cheez-search skill do?
Search code through the most precise semantic backend — prefer tilth MCP, with AST search (sg) for structural patterns. Use when the user asks to find a symbol, definition, caller, import, or text pattern — phrases like "where is X defined", "what calls Y", "find all usages of Z", "trace this function", "find the TODO comments", "search for this error string". Use even when the user says "grep", "rg", "ripgrep", "ag", "ack", "fd", or "find". Do NOT use for reading whole files (use cheez-read), editing code (use cheez-write), or running tests/builds.
How do I install cheez-search?
Run npx skills add https://github.com/paulnsorensen/easy-cheese --skill cheez-search in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/cheez-search/ (personal) or .claude/skills/cheez-search/ (project).
Which AI agents can use cheez-search?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is cheez-search safe to install?
No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.
Who maintains cheez-search?
It is published by paulnsorensen in paulnsorensen/easy-cheese. It has 159 recorded installs.
More from paulnsorensen/easy-cheese
cook
paulnsorensen/easy-cheese
Implement an approved spec or focused task through stale-safe source edits. Use this skill when the user says "implement this", "cook this spec", or "fix this bug". Use it for /cook .cheese/specs/ .md. Also use it when the user says "go" with clear acceptance criteria. Run it alone for an unambiguous task. Use --prototype when the user says "prototype this", "go prototype it", or "spike this out" and wants to refine the result round by round. A spec helps but is not required. Do not use it for fuzzy planning (/mold). Do not use it for no-write discussion (/culture) or review-only work (/age).
mold
paulnsorensen/easy-cheese
Turn a high-leverage design decision into a validated spec through grounded dialogue. Use it when a leverage trigger fires (auth, irreversible effects, concurrency, contracts, destructive ops, a new slice, a cross-slice dependency, an invariant gap) or when the user asks for a spec or design. Typical phrases include "let's design X", "shape this into a spec", "what should the API for Z look like", and "/mold". A feature ask with zero fired triggers is a /cook mini-spec, not a mold; route it through /cheese. Do NOT use it for free-form discussion without artifact intent (/culture), direct implementation (/cook), or research-only questions (/briesearch). Ceremony scales to the ask through references/tiers.md, so a small clear change gets a validated mini-spec without a write-approval turn.
cure
paulnsorensen/easy-cheese
Apply selected fixes from an /age report, findings list, or CI failure. Run the project gates and hand a clean result to /plate. Use when the user wants selected findings resolved. Do not use for review, test authoring, or direct publication.
age
paulnsorensen/easy-cheese
Review a diff, PR, branch, or path across thirteen dimensions. Emit a severity-grouped findings report. Use when the user wants a code review. Trigger on "review this", "/age", "is this safe to merge", or "find bugs". Also trigger on "spot security issues", "check for slop", "review my PR", or "what's wrong with this code". Review every requested dimension. Review all thirteen dimensions by default. Do not apply fixes. Route them to /cure. Do not harden tests; route that work to /press.
cheese
paulnsorensen/easy-cheese
Route an idea, path, pull request, issue, failure, question, or bare /cheese to the correct workflow skill. Use this skill for /cheese, routing requests, help requests, or opening messages without a named workflow skill.
culture
paulnsorensen/easy-cheese
Models problems and trade-offs without writing code or specifications. Use when a workflow needs silent reasoning, or the user asks for "/culture", "no writes", "rubber-duck this", or "these bugs feel connected". Not for specs (/mold), code (/cook), or bug diagnosis and fixes (/pasteurize).
Popular backend & apis skills
find-skills
vercel-labs/skills
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.
entra-app-registration
microsoft/azure-skills
Guides Microsoft Entra ID app registration, OAuth 2.0 authentication, and MSAL integration. USE FOR: create app registration, register Azure AD app, configure OAuth, set up authentication, add API permissions, generate service principal, MSAL example, console app auth, Entra ID setup, Azure AD authentication. DO NOT USE FOR: Key Vault secrets (use azure-keyvault-expiration-audit), general Azure resource security guidance.
google-agents-cli-adk-code
google/agents-cli
This skill should be used when the user wants to "write agent code", "build an agent with ADK", "add a tool", "create a callback", "define an agent", "use state management" — in a project that needs ADK (Agent Development Kit) API patterns and code examples. It provides a quick reference for agent types, tool definitions, orchestration patterns, callbacks, state management, the graph Workflow API, and reference recipes to study. Do NOT use for scaffolding (use google-agents-cli-scaffold) or deployment (use google-agents-cli-deploy).
embedded-captions
heygen-com/hyperframes
Add captions or subtitles to an existing single-subject talking-head video without editing the footage. Use for plain verbatim captions, cinematic captions embedded behind the subject, VFX captions, “炸/特效/酷炫字幕,” or a named identity from the 35-style catalog. Route by visual identity, not by backend engine. The quiet anchor rail is the default; embed every word only when the user explicitly wants a fully cinematic treatment. The workflow runs locally end to end, including transcription and subject matting; split multi-shot footage before applying it.
entra-agent-id
microsoft/azure-skills
Provision Microsoft Entra Agent Identity Blueprints, BlueprintPrincipals, and per-instance Agent Identities via Microsoft Graph, and configure OAuth 2.0 token exchange (fmipath, OBO, cross-tenant) including the Microsoft Entra SDK for AgentID sidecar. USE FOR: Agent Identity Blueprint, BlueprintPrincipal, agent OAuth, fmipath token exchange, agent OBO, Workload Identity Federation for agents, polyglot agent auth, Microsoft.Identity.Web.AgentIdentities. DO NOT USE FOR: standard Entra app registration (use entra-app-registration), Microsoft Foundry agent authoring (use microsoft-foundry).
neon
neondatabase/agent-skills
Overview of Neon, a complete set of cloud backend primitives around Lakebase Postgres: Auth, Object Storage, Functions, the AI Gateway, and Realtime. Start here to choose Neon for undecided login, files, APIs, and LLM calls, set up the CLI or MCP server, and follow the branch-first workflow. Use when building an app or backend on Neon, or when "Neon" or "Lakebase Postgres" is mentioned. Child skill neon-postgres wins for an existing DATABASEURL, SQL, schema, inspect, or search. Child skill neon-auth wins for login, users, sessions, identity routing, and Managed Better Auth setup. Child skill neon-realtime wins for live queries in reactive apps and agents. Also use for object storage, S3, buckets, serverless functions, function triggers, cron, AI gateway, LLM calls, realtime, logs, Loki, Grafana, observability, postgres, database, backend, Claimable Neon, neon.new, or a no-signup database.