claude

testvdb

通过自然语言契约逆向工程从官方文档提取结构化约束,结合多 Agent 辩论机制在 Docker 沙箱中自动发现向量数据库的合规性缺陷。

0 on GitHubby TestVDBv2.4.0MITupdated 2 months agotesting
/

Plugin testvdb from marketplace testvdb in yihui504/TestVDB. Commands read from the repo's manifest — 1 month ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use and it starts local MCP servers. The exact commands are listed below; read them before installing.

What it installs

Skills (4)

  • contract-schema

    TestVDB structured-contract JSON Schema reference. Auto-loaded when the Contract Formalizer agent or related agents need the contract format.

  • defect-taxonomy

    TestVDB four-type defect taxonomy reference. Auto-loaded when Judge or Attack agents need to determine defect types.

  • docker-templates

    TestVDB Docker container template reference. Auto-loaded when the Executor agent needs to start target vector-database containers.

  • pipeline

    TestVDB defect-mining pipeline SOP. Auto-loaded when the Orchestrator coordinates the defect-mining pipeline.

Slash commands (4)

  • /testvdb:contract

    Generate/refresh documentation knowledge and the structured contract for a specified DB version

  • /testvdb:intel

    Collect historical Issue/Commit intelligence for a specified DB and build the threat model

  • /testvdb:mine

    Launch the vector-database automated defect-mining pipeline

  • /testvdb:resume

    Discover unfinished mining runs and resume them

Subagents (17)

  • _target_api_reference
  • api-template-formalizer

    Distills focused API syntax templates (request bodies / response structures) from raw_knowledge.json, for attack agents to consume on demand.

  • attack-boundary

    Boundary attack agent — focuses on test generation for parameter boundary-value violations.

  • attack-semantic

    Semantic attack agent — focuses on test generation for behavioral-contract violations, error diagnostics quality, and search semantic correctness.

  • attack-state

    State attack agent — focuses on test generation for data-consistency, concurrent-operation, and state-transition violations.

  • bug-shape-extractor

    Bug Shape extraction agent — three-way classifies historical issues and extracts root-cause patterns and developer cognition boundaries.

  • chain-auditor

    Evidence-chain audit agent (dedicated) — reads only the evidence-chain files, runs the completeness/consistency/self-consistency triple check and the multi-perspective aggregation, and produces the final truth verdict. Performs no forensics.

  • contract-formalizer

    Converts the raw API knowledge document into a structured, machine-readable contract JSON.

  • docker-executor

    Docker sandbox execution agent — runs attack scripts in isolated containers and collects results.

  • evidence-builder

    Evidence-chain building agent — collects documentation verification, execution-evidence review, and source-code forensics for a single candidate defect, writing the evidence-chain file. Makes no truth judgment.

  • issue-miner

    Historical issue mining agent — crawls the target repository's issues and merged PRs to build the raw defect corpus.

  • knowledge-extractor

    Extracts API knowledge and constraint information for the target vector database from official documentation.

  • orchestrator-lifecycle

    Orchestrator lifecycle management — error handling, context-compaction protection, progress visibility, multi-DB parallelism.

  • orchestrator

    TestVDB defect-mining pipeline chief orchestrator. Coordinates all 16 agents through the full flow from strategic intelligence collection to defect reporting.

  • reporter-mre

    MRE script generation agent — produces self-contained minimal reproduction scripts for confirmed defects.

  • reporter

    Defect-report generation agent — produces standardized Markdown reports for confirmed defects (MRE scripts are generated separately by reporter-mre).

  • threat-modeler

    Threat-model construction agent — builds the Threat Model and Developer Cognitive Blindspot model from historical defect data.

MCP servers (1)

  • github

    runs: npx -y @modelcontextprotocol/server-github

Hooks (7)

  • SessionStart

    python ${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py ${CLAUDE_PLUGIN_ROOT}/scripts/preflight.py

  • PostToolUse

    python ${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py ${CLAUDE_PLUGIN_ROOT}/scripts/log_execution.py

  • PostToolUse

    python ${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py ${CLAUDE_PLUGIN_ROOT}/scripts/write_location_check.py

  • PostToolUseFailure

    python ${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py ${CLAUDE_PLUGIN_ROOT}/scripts/retry_policy.py

  • Stop

    python ${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py ${CLAUDE_PLUGIN_ROOT}/scripts/hooks/pipeline_gate.py

  • PreCompact

    python ${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py ${CLAUDE_PLUGIN_ROOT}/scripts/precompact_save.py

  • PostCompact

    python ${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py ${CLAUDE_PLUGIN_ROOT}/scripts/postcompact_verify.py

View source on GitHub

Similar plugins

Collection of example skills demonstrating various capabilities including skill creation, MCP building, visual design, algorithmic art, internal communications, web testing, artifact building, Slack GIFs, and theme styling

19 skills

anthropics/skills 193,780

Comprehensive PR review agents specializing in comments, tests, error handling, type design, code quality, and code simplification

1 command · 6 agents

anthropics/claude-code 155,476

storybook

claudecodex

Build, preview, and test UI components with Storybook.

7 skills

storybookjs/storybook 91,552

Agentic QE (external: maintained in proffesor-for-testing/agentic-qe, pinned to v3.15.17): a quality-engineering fleet with 11 QE agents, 9 skills, 9 commands, the agentic-qe MCP server pinned to the plugin version, and the aqe-mod guard that refuses destructive operations on AQE learning data

9 skills · 9 commands · 11 agents · 1 MCP server

ruvnet/ruflo 81,589

Automates browser interactions for web testing, form filling, screenshots, and data extraction

12 skills

vercel-labs/agent-browser 50,373

Maintain testvdb?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex pluginslisted