New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Trusty Search

by bobmatnyc in bobmatnyc/claude-mpm

Hybrid code search (BM25 + vector + KG) with RRF fusion. An agent skill from bobmatnyc/claude-mpm.

Apache-2.0Auto-check passedAI & LLM Engineering

Install Trusty Search

skills CLI
$ npx skills add bobmatnyc/claude-mpm --skill trusty-search -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install bobmatnyc/claude-mpm trusty-search --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/bobmatnyc/claude-mpm.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/trusty-search .claude/skills/trusty-search && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
trusty-search
GitHub stars
156
Token cost
~1.5k tokens
SKILL.md length
364 words
Files
2
Skills in repo
51
Repo updated
First seen
Licence
Apache-2.0

At a glance

Hybrid code search (BM25 + vector + KG) with RRF fusion. An agent skill from bobmatnyc/claude-mpm.

  • Tasks that involve Vector databases
  • SKILL.md covers When to Use, Architecture, Query Type Routing and Tool Reference, plus 6 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
  • Tasks that involve Codebase onboarding

What it does

Trusty Search is an agent skill from bobmatnyc/claude-mpm. Hybrid code search (BM25 + vector + KG) with RRF fusion. Single daemon serves multiple named indexes. Replaces mcp-vector-search.

Its SKILL.md is about 1.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 1 other file (for example `metadata.json`). Compatibility notes: claude-code

It sits in AI & LLM Engineering, covering Vector databases and Codebase onboarding. It works with Model Context Protocol. The repository describes itself as: Claude Multi-Agent Project Manager — multi-channel orchestration, GitHub-first SDK mode, and plugin system for Claude. The licence is Apache-2.0.

When your agent uses it

  • Tasks that involve Vector databases
  • Tasks that involve Codebase onboarding

Example prompts

  • “/trusty-search”

Requirements

  • Compatibility (from SKILL.md): claude-code

What it can do on your machine

Read from SKILL.md and the folder at commit 25203d3. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

  • Compatibility

    claude-code

    From compatibility in the SKILL.md frontmatter.

Context cost

Trusty Search loads about 1.5k tokens when it runs. Until then it costs about 36 tokens; SKILL.md has 364 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~36
When it runs · the whole SKILL.md, loaded when a task matches
~1.5k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from bobmatnyc/claude-mpm at commit 25203d3, republished under its Apache-2.0 licence (© bobmatnyc). 364 words, ~1,480 tokens.

Download SKILL.mdSave it as .claude/skills/trusty-search/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
trusty-search
description
Hybrid code search (BM25 + vector + KG) with RRF fusion. Single daemon serves multiple named indexes. Replaces mcp-vector-search.
compatibility
claude-code
user-invocable
false
disable-model-invocation
false
license
Apache-2.0

Hybrid semantic + lexical + relational code search. Replaces mcp__mcp-vector-search__* tools.

When to Use

  • Locate code: "Where is X defined?" / "How is X used?"
  • Explore concepts: "How does the migration system work?"
  • Find tech debt: "Where are the TODOs / FIXMEs?"
  • Q&A over codebase: Use chat for synthesized answers with citations.

Do NOT use for: persistent memory (use trusty-memory), edits (use Edit), or shell-level file listing.

Architecture

  • Single daemon at http://127.0.0.1:7878 serves all projects.
  • Named indexes — one per project: claude-mpm, cto, duetto, open-mpm, rust-dre, etc.
  • Hybrid retrieval:
    • BM25 — lexical/keyword precision
    • HNSW — vector semantic similarity
    • KG — knowledge graph relational signals
    • Fused with Reciprocal Rank Fusion (RRF)

Query Type Routing

search_code accepts a query_type hint that biases the fusion weights. Always specify it.

query_typeUse ForBias
Definition"Where is run_pending_migrations defined?"BM25 precision
Usage"How is the planner agent invoked?"Vector similarity
Conceptual"How does the migration system work?"Balanced BM25+vector+KG
BugDebt"Show me the TODOs / hacks"Keyword match

If unspecified, the daemon auto-routes — but explicit is better.

Tool Reference

search_code — primary tool.

mcp__trusty-search__search_code(
  query="migration registry",
  index="claude-mpm",
  query_type="Definition",
  limit=10
)

chat — LLM-powered Q&A over the index with citations.

mcp__trusty-search__chat(
  index="claude-mpm",
  question="How does the hook dispatcher route PreToolUse events?"
)
Index Management
mcp__trusty-search__list_indexes()
mcp__trusty-search__create_index(name="new-project", root="/abs/path")
mcp__trusty-search__delete_index(name="old-project")
mcp__trusty-search__index_status(index="claude-mpm")
mcp__trusty-search__reindex(index="claude-mpm")
mcp__trusty-search__index_file(index="claude-mpm", path="src/foo.py")
mcp__trusty-search__remove_file(index="claude-mpm", path="src/old.py")
Health
mcp__trusty-search__search_health()
# Returns daemon liveness + uptime.

When to Use search_code vs chat

NeedTool
List of matching files/symbolssearch_code
Quick lookup, then read sourcesearch_code
Synthesized explanation across many fileschat
Citations and a narrative answerchat
Bulk grep replacementsearch_code with query_type="Definition"

search_code is cheap; chat is more expensive. Prefer search_code + targeted Read when feasible.

Show full SKILL.md (132 more words)Show less

Index Health Workflow

# 1. Confirm daemon up
search_health()

# 2. Confirm index registered and healthy
index_status(index="claude-mpm")

# 3. If stale or recently changed many files:
reindex(index="claude-mpm")

# 4. For single-file invalidation after edit:
index_file(index="claude-mpm", path="src/claude_mpm/foo.py")

If search_health fails: the daemon is not running. Start it externally (trusty-search serve). The MCP transport assumes the daemon is already serving.

Example Calls

Definition lookup
search_code(
  query="run_pending_migrations",
  index="claude-mpm",
  query_type="Definition",
  limit=5
)
Conceptual exploration
search_code(
  query="how hooks get dispatched after tool use",
  index="claude-mpm",
  query_type="Conceptual",
  limit=15
)
Tech debt audit
search_code(
  query="TODO FIXME XXX HACK",
  index="claude-mpm",
  query_type="BugDebt",
  limit=50
)
Q&A
chat(
  index="claude-mpm",
  question="What invariants does the migration registry enforce on migration IDs?"
)

Integration with trusty-memory

Search → understand → remember. After investigation, persist learnings.

# 1. Find code
search_code(query="model_tier_hook", index="claude-mpm", query_type="Definition")

# 2. Read and understand the code (use Read tool)

# 3. Persist the learning
mcp__trusty-memory__memory_remember(
  content="model_tier_hook enforces planner→claude-opus-4-7 routing; config at ~/.claude-mpm/config/configuration.yaml models.planning",
  room="Backend",
  tags=["hooks", "model-tier", "planner"],
  importance=0.8
)

# 4. Optionally assert relationship in KG
mcp__trusty-memory__kg_assert(
  subject="model_tier_hook",
  predicate="enforces",
  object="planner-model-routing"
)
OldNew
mcp__mcp-vector-search__search_codemcp__trusty-search__search_code
mcp__mcp-vector-search__index_projectmcp__trusty-search__create_index + reindex
mcp__mcp-vector-search__get_project_statusmcp__trusty-search__index_status
mcp__mcp-vector-search__search_similarsearch_code with query_type="Usage"
mcp__mcp-vector-search__search_contextsearch_code with query_type="Conceptual" or chat
mcp__mcp-vector-search__search_hybridsearch_code (hybrid is default)

Best Practices

  • Always pass index: the daemon hosts many projects; omitting index is ambiguous.
  • Specify query_type: it materially improves ranking.
  • Start narrow: limit=10 for search_code, then expand if needed.
  • Reindex sparingly: full reindex is expensive — prefer index_file for single-file updates.
  • Verify before claiming: check search_health if results seem stale or empty.
  • Pair with memory: search to find, memory to remember.

© bobmatnyc, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file in .claude/skills/trusty-search of bobmatnyc/claude-mpm.

  • SKILL.md
  • metadata.json

Open the folder on GitHubat commit 25203d3

Compare with similar skills

Trusty Search next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Trusty Search compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Trusty Search this skillbobmatnyc/claude-mpm156—~1.5kAutomated safety check: PassApache-2.0
Codebase Explorationgiancarloerra/SocratiCode3.3k1 repos~1.5kAutomated safety check: PassAGPL-3.0
Codebase Managementgiancarloerra/SocratiCode3.3k1 repos~1.8kAutomated safety check: PassAGPL-3.0
Flowflow Spacesmirkobozzetto/flowflow171—~1kAutomated safety check: PassEUPL-1.2
Pgvector Semantic Searchtimescale/pg-aiguide1.9k—~3.8kAutomated safety check: PassApache-2.0
Postgres Hybrid Text Searchtimescale/pg-aiguide1.9k—~3.1kAutomated safety check: PassApache-2.0

Similar skills

  • Codebase Exploration

    giancarloerra/SocratiCode

    Explore and understand codebases using SocratiCode semantic search, dependency graphs, and context artifacts.

    3.3k GitHub starsUsed in 1 repo~1.5k tokens
    DatabasesAuto-check passed

  • Codebase Management

    giancarloerra/SocratiCode

    Set up, index, and manage SocratiCode codebase indexing. An agent skill from giancarloerra/SocratiCode.

    3.3k GitHub starsUsed in 1 repo~1.8k tokens
    AI & LLM EngineeringAuto-check passed

  • Flowflow Spaces

    mirkobozzetto/flowflow

    Use FlowFlow spaces safely through their scoped MCP server. An agent skill from mirkobozzetto/flowflow.

    171 GitHub stars~1k tokensUpdated 2 days ago
    AI & LLM EngineeringAuto-check passed

  • Pgvector Semantic Search

    timescale/pg-aiguide

    A skill your agent uses for setting up vector similarity search with pgvector for AI/ML embeddings, RAG applications, or semantic search.

    1.9k GitHub stars~3.8k tokensUpdated 3 days ago
    AI & LLM EngineeringAuto-check passed

  • Postgres Hybrid Text Search

    timescale/pg-aiguide

    A skill your agent uses to implement hybrid search combining BM25 keyword search with semantic vector search using Reciprocal Rank Fusion (RRF).

    1.9k GitHub stars~3.1k tokensUpdated 3 days ago
    AI & LLM EngineeringAuto-check passed

  • Vera

    VeraTools/vera

    Code search over the current repository. An agent skill from VeraTools/vera.

    119 GitHub stars~1.2k tokensUpdated yesterday
    AI & LLM EngineeringAuto-check passed

More from bobmatnyc/claude-mpm

All 51 skills in this repo

  • Build MCP Server

    bobmatnyc/claude-mpm

    Create high-quality MCP servers that enable LLMs to effectively interact with external services.

    156 GitHub stars~2k tokensUpdated 1 mo ago
    Auto-check passed

  • Env Manager

    bobmatnyc/claude-mpm

    Environment variable validation, synchronization, and management across local development, CI/CD, and deployment platforms

    156 GitHub stars~2k tokensUpdated 1 mo ago
    Auto-check: notes

  • Session Analyzer

    bobmatnyc/claude-mpm

    Debug and teach agentic coding: a deterministic-first session timeline + cost report, with optional narrative polish and a standalone JSX visualiser.

    156 GitHub stars~2.2k tokensUpdated 1 mo ago
    Auto-check passed

  • Software Patterns

    bobmatnyc/claude-mpm

    Decision framework for architectural patterns including DI, SOA, Repository, Domain Events, Circuit Breaker, and Anti-Corruption Layer.

    156 GitHub stars~1.8k tokensUpdated 1 mo ago
    Auto-check passed

  • Verification Before Completion

    bobmatnyc/claude-mpm

    Run verification commands and confirm output before claiming success

    156 GitHub starsUsed in 2 repos~1k tokens
    Auto-check passed

  • Dependency Audit

    bobmatnyc/claude-mpm

    Dependency audit and cleanup workflow for maintaining healthy project dependencies.

    156 GitHub stars~3.5k tokensUpdated 1 mo ago
    Auto-check passed

Works with

Categories

Questions about Trusty Search

What does Trusty Search do?

Hybrid code search (BM25 + vector + KG) with RRF fusion. An agent skill from bobmatnyc/claude-mpm. Trusty Search is an agent skill from bobmatnyc/claude-mpm. Hybrid code search (BM25 + vector + KG) with RRF fusion.

When should I use Trusty Search?

Trusty Search fits situations like: tasks that involve Vector databases; tasks that involve Codebase onboarding.

How do I install Trusty Search in Claude Code?

Run `npx skills add bobmatnyc/claude-mpm --skill trusty-search -a claude-code`. Or copy the skill folder (.claude/skills/trusty-search in bobmatnyc/claude-mpm) into .claude/skills/trusty-search in your project. Claude Code loads it when a task matches its description.

How do I install Trusty Search in Codex?

Run `npx skills add bobmatnyc/claude-mpm --skill trusty-search -a codex`. Or copy the skill folder (.claude/skills/trusty-search in bobmatnyc/claude-mpm) into .agents/skills/trusty-search in your project. Codex loads it when a task matches its description.

Can I use Trusty Search in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add bobmatnyc/claude-mpm --skill trusty-search -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/trusty-search, .gemini/skills/trusty-search, .github/skills/trusty-search and .opencode/skills/trusty-search in your project.

What does Trusty Search need to run?

SKILL.md names no scripts, command-line tools or credentials: Trusty Search is instructions for the agent only. Compatibility (from SKILL.md): claude-code.

Does Trusty Search access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Trusty Search safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Trusty Search use?

Trusty Search is published under the Apache-2.0 licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Trusty Search use?

About 1.5k tokens (SKILL.md is roughly 5.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Trusty Search?

Skills that share tags, products or a category with Trusty Search: Codebase Exploration (giancarloerra/SocratiCode, 3.3k stars), Codebase Management (giancarloerra/SocratiCode, 3.3k stars), Flowflow Spaces (mirkobozzetto/flowflow, 171 stars) and Pgvector Semantic Search (timescale/pg-aiguide, 1.9k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Trusty Search?

bobmatnyc (a GitHub user) maintains it in bobmatnyc/claude-mpm, which has 156 GitHub stars. The repository holds 51 skills in this directory. The repository was last updated on August 31, 2026.

Source: bobmatnyc/claude-mpm on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.