Codesteward
Popular repositories Loading
-
codesteward-audit-proxy
codesteward-audit-proxy PublicTransparent reverse proxy for LLM APIs (Anthropic, OpenAI, Gemini) that captures every agent request and response to ClickHouse for audit, analysis, and summarization — zero agent configuration req…
Go 1
-
codesteward
codesteward PublicStructural code graph server for AI agents — parses repositories into a Neo4j graph via tree-sitter and exposes it as an MCP tool interface with lexical, referential, semantic, and dependency queries.
Python
-
codesteward-taint
codesteward-taint PublicTaint-flow analysis engine for the Codesteward platform — traces untrusted input to dangerous sinks across your codebase using three-level analysis: call graph traversal, intra-procedural propagati…
-
codesteward-audit-proxy-plugins
codesteward-audit-proxy-plugins PublicIDE plugins (VSCode + JetBrains) that configure the Codesteward audit proxy for AI coding agents, automatically injecting identity metadata as X-Audit-* headers for structured audit logging.
Kotlin
-
codesteward-dashboard
codesteward-dashboard PublicUnified web dashboard for the Codesteward platform — monitor AI coding agent activity, explore code graphs with a force-directed visualizer, and review taint-flow security findings. Built with Nuxt…
Vue
-
codesteward-session-summarizer
codesteward-session-summarizer PublicBackground service that summarizes AI coding sessions from ClickHouse audit events using a local LLM via Ollama.
Python
Repositories
- codesteward Public
Structural code graph server for AI agents — parses repositories into a Neo4j graph via tree-sitter and exposes it as an MCP tool interface with lexical, referential, semantic, and dependency queries.
Codesteward/codesteward’s past year of commit activity - codesteward-audit-proxy Public
Transparent reverse proxy for LLM APIs (Anthropic, OpenAI, Gemini) that captures every agent request and response to ClickHouse for audit, analysis, and summarization — zero agent configuration required.
Codesteward/codesteward-audit-proxy’s past year of commit activity - codesteward-session-summarizer Public
Background service that summarizes AI coding sessions from ClickHouse audit events using a local LLM via Ollama.
Codesteward/codesteward-session-summarizer’s past year of commit activity - codesteward-prompt-evaluator Public
Batch service that scores the quality of outputs from codesteward-session-summarizer using a strong LLM (Claude, GPT-4o) and automatically optimizes prompts through an A/B validation loop. Communicates with the summarizer exclusively through shared ClickHouse tables — no runtime coupling between the two services.
Codesteward/codesteward-prompt-evaluator’s past year of commit activity - codesteward-dashboard Public
Unified web dashboard for the Codesteward platform — monitor AI coding agent activity, explore code graphs with a force-directed visualizer, and review taint-flow security findings. Built with Nuxt 4, Vue 3, and ECharts. Each backend module is independently optional.
Codesteward/codesteward-dashboard’s past year of commit activity - codesteward-audit-proxy-plugins Public
IDE plugins (VSCode + JetBrains) that configure the Codesteward audit proxy for AI coding agents, automatically injecting identity metadata as X-Audit-* headers for structured audit logging.
Codesteward/codesteward-audit-proxy-plugins’s past year of commit activity - codesteward-taint Public
Taint-flow analysis engine for the Codesteward platform — traces untrusted input to dangerous sinks across your codebase using three-level analysis: call graph traversal, intra-procedural propagation, and path-sensitive CFG analysis.
Codesteward/codesteward-taint’s past year of commit activity
Top languages
Loading…
Most used topics
Loading…