maven-code-search — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited maven-code-search (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Use the maven-decoder MCP server before guessing about third-party Java dependencies. It can search the user's installed Maven repository, inspect jars, prefer source jars when available, and decompile bytecode when source is missing.
This MCP only sees artifacts already present in the local Maven cache. If the needed artifact is missing, ask the user to build the project or run a dependency resolution command such as mvn dependency:go-offline, then retry the MCP search.
group_id, artifact_id, and version. Read pom.xml, build.gradle, lockfiles, or existing imports if the user did not provide them.Use list_artifacts to discover installed artifacts or filter by partial group_id, artifact_id, or version.
Use get_version_info when the user asks which versions of an artifact are installed.
Use search_classes when the user knows a class name, simple type name, wildcard, or package but not the artifact that contains it.
Use extract_class_info when the user needs constructors, fields, annotations, or method signatures for classes in a known artifact. If sources are missing, this tool still returns bytecode-backed fields and methods through the MCP, so do not shell out to javap for the same information.
Use extract_source_code when the user needs implementation details. Set prefer_sources to true unless there is a specific reason to force decompilation.
Use extract_jar_resource when the user needs files embedded in the jar, such as .proto files, service descriptors, or Maven metadata. Use analyze_jar first when you need to discover resource paths. Do not shell out to jar tf just to list or read jar entries.
Use extract_method_info for targeted method inspection instead of pulling a large class.
Use get_dependencies for direct POM dependencies and get_dependency_tree for transitive dependency chains.
Use find_dependents when the user asks "what installed artifacts depend on X?"
Use compare_versions when the user asks what changed between two installed versions of the same artifact.
Use analyze_jar or analyze_jar_structure when the user asks about jar metadata, manifests, packages, resources, services, or high-level structure.
Use find_usage_examples when the user asks how a class or method is used in installed test jars.
For "Where is ObjectMapper installed?", call search_classes with class_name: "ObjectMapper" and inspect the returned artifact coordinates.
For "Show me the methods on RestTemplate", first resolve the artifact/version if needed, then call extract_class_info with the fully qualified class name or a precise class pattern.
For "How does this library implement retry?", search for likely classes by package or class name, then use extract_method_info or extract_source_code on the most relevant classes.
For generated protobuf classes with no sources jar, call extract_class_info on the generated class to recover parsed methods and fields. If the jar includes .proto resources or descriptor metadata, use analyze_jar and extract_jar_resource to inspect those resources through the MCP.
For "Why do I have two versions?", use get_dependency_tree, find_dependents, and compare_versions as needed. Tie the answer back to the project's declared dependencies.
Keep responses grounded in MCP results. Include:
Do not paste huge decompiled files into chat. Summarize the relevant behavior and include only small snippets when they are necessary to explain the answer.
If the maven-decoder MCP tools are not available, tell the user the skill requires the Maven Decoder MCP server to be installed and configured in their MCP-capable client. A common stdio setup is:
{
"maven-decoder": {
"command": "uvx",
"args": ["maven-decoder-mcp"]
}
}~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.