Mukunda Rao Katta
@mukundakatta
AI/ML Engineer | Building agentic workflows & LLM integrations | Anthropic Certified | Open Source |
95
Published Tools
13
Total Stars
39
Weekly Downloads
43
GitHub Followers
1,396
Public Repos
100/100
Avg Security
Published Tools
37 MCP Servers32 Skills26 Agentsacross 8 categoriesMukunda Katta <[email protected]>
Linter for MCP (Model Context Protocol) config files used by Claude Desktop, Cursor, Cline, Windsurf, and Zed. CLI + library API.
...moreMukundaKatta
Structured-output enforcer: extract and validate JSON from messy LLM text.
MukundaKatta
Token-aware message truncation: fit a chat history into your model's context budget.
MukundaKatta
Network-egress firewall for agent tools: check URLs against a declarative policy.
MukundaKatta
Validate tool-call args before execution. Returns LLM-friendly retry hints.
mukundakatta
Validate MCP (Model Context Protocol) config files for Claude, Cursor, Cline, Windsurf, and Zed. CLI + GitHub Action with SARIF output.
...moreMukundaKatta
A small, dependency-free Node.js CLI that watches a fleet of MCP (Model Context Protocol) servers and reports health, latency, and uptime.
...moreMukundaKatta
Minimal Python toolkit for creating MCP servers from plain functions using decorators
MukundaKatta
A small, dependency-free Python template for building Model Context Protocol (MCP) servers.
MukundaKatta
Developer toolkit for building Model Context Protocol servers and clients
MukundaKatta
MCPForge — Build, test, and deploy Model Context Protocol servers in minutes. Scaffolding, hot-reload dev server, and registry ublishing for the MCP ecosystem.
...moremukundakatta
Inject, parse, and strip [N] citation markers in RAG outputs.
mukundakatta
MCP server for RAG retrieval evaluation metrics: Recall@k, Hit@k, MRR, NDCG@k. Helps LLMs score retrieval quality.
mukundakatta
MCP server for token-budget-aware text truncation. Helps an LLM fit text into a context window mid-task.
Mukunda Katta <[email protected]>
Lint MCP config files for Claude Desktop, Claude Code, Cursor, Cline, Windsurf, and Zed. Stdlib-only Python port of @mukundakatta/mcpcheck.
...moremukundakatta
MCP server: convert times between IANA timezones, list zones, get current offset.
mukundakatta
MCP server for RAG drift diagnostics. Interpret drift scores, recommend thresholds, and explain the five drift dimensions.
...moreStructured output for any LLM call. Validate-and-retry loop for JSON responses; BYO LLM and validator. Python port of @mukundakatta/agentcast.
...moreFit your messages into the LLM context window. Token-aware truncation with multiple strategies, pluggable tokenizers. Python port of @mukundakatta/agentfit.
...moreIdempotency keys for LLM agent retries. Deterministic content-derived keys (sha256-hex, UUIDv5, scoped). Pairs with cachebench miss-aware retry. Zero runtime deps.
...moreSnapshot tests for AI tool-call traces: diff, validate, and catch silent regressions.
MukundaKatta
Wire-level prompt introspection for LLM SDK calls. See exactly what was sent, with credentials redacted by default. Anthropic, OpenAI, any httpx-based client.
...moreValidate LLM-generated tool args before execution. Wraps tool functions with arg validation, raises ToolArgError with LLM-friendly retry hint. Python port of @mukundakatta/agentvet.
...moreZero-dependency eval harness for LLM and agent regression testing. Scores outputs with exact, contains, regex, JSON, citation, and token-F1 checks.
...moreAudit and fix Anthropic prompt caching on AWS Bedrock through any abstraction stack.
Low-level Python ergonomics for AWS Bedrock + Anthropic Claude: retries, cost ledger, streaming-error normalization.
MukundaKatta
Prompt-cache observability for LLM APIs. Per-call hit ratios, cost saved, regression alerts, miss-aware retry. Anthropic + OpenAI + Bedrock.
...moreLinter for Claude Code slash-command files (.claude/commands/*.md). Validates YAML frontmatter, allowed-tools shape, description quality, and flags hardcoded secrets.
...moreLinter for Claude Code hooks configuration (the 'hooks' block of settings.json). Validates event names, matcher shape, command entries, and flags dangerous commands or hardcoded secrets.
...moreLinter for Claude Code SKILL.md files. Validates YAML frontmatter, required fields, description length, and common secret patterns.
...moreContext engineering toolkit for ranking, packing, and risk-scanning RAG context. Python port of @mukundakatta/context-forge.
...moreMukunda Katta
Embedding pipeline ops + drift detection for production RAG: index manifests, version assertions, neighbor-stability eval, Drift-Adapter for in-place model migrations.
...moreDetect flaky LLM eval cases across repeated runs. Pass-rate + standard-deviation per case, with per-case severity. Python port of @mukundakatta/eval-flake-detector.
...moreEstimate hallucination risk in LLM answers from uncertainty language, unsupported specifics, citations, and context coverage. Python port of @mukundakatta/hallucination-risk-meter.
...moreEstimate LLM request cost and enforce per-request or per-session budgets. Python port of @mukundakatta/llm-cost-guard.
Stable canonical sha256 hash of LLM request/message structures. Recursive key-sorted JSON canonicalization with per-provider presets that drop noise fields. For cache keys and idempotency. Zero runtime deps.
...moreSanitize LLM outputs before HTML, SQL, shell, or markdown sinks. Python port of @mukundakatta/llm-output-sanitizer.
Tiny schema validator for structured LLM responses. Python port of @mukundakatta/llm-response-schema-lite.
Exponential backoff with full jitter for LLM API calls. Sync + async. Built-in retryable-code presets for Anthropic, OpenAI, Bedrock, Gemini. Zero runtime deps.
...moreMukunda Katta <[email protected]>
Sample LLM traces by risk, errors, latency, and deterministic ids. Python port of @mukundakatta/llm-trace-sampler.
Parse LLM API response logs (Anthropic, OpenAI, Google) and generate token / cost reports. No framework adoption required.
...moreMukundaKatta
Fleet-level batch dispatcher for LLM APIs. Pool requests across coroutines, route to provider Batch APIs, save 50% on cost without rewriting your agent loops.
...moreMukunda Katta
Sub-millisecond PII redaction for prompts before they reach an LLM.
Reliable CSV parsing + generation. Handles quotes, embedded commas, BOMs, CRLF.
Character-precise unified diff + patch application + patch parsing.
Convert HTML to Markdown or strip to plain text. For web-scraping agents.
Run JMESPath queries against JSON. Pure JS, no native binary needed.
Parse JSON5 (comments, trailing commas, unquoted keys) and emit strict JSON.
Trustworthy JS regex testing. Real offsets, named groups, safe against zero-width loops.
Safe shell argument escaping for bash, cmd.exe, and PowerShell.
Deterministic SQL formatting across 19 dialects (postgres, mysql, snowflake, bigquery, etc.).
Parse, format, and convert configs across TOML, YAML, and JSON.
Streaming JSON parser that yields partial valid trees as tokens arrive. For LLM tool calls, structured outputs, and partial recovery.
...moreHeuristic quality metrics for RAG retrieval and grounded answers. Python port of @mukundakatta/rag-quality-kit.
Find stale RAG chunks by age, version, and freshness requirements. Python port of @mukundakatta/rag-staleness-auditor.
Five-dimensional drift detection for production RAG systems. Rust core, Python bindings.
Stable semantic cache keys for LLM requests. Invariant to whitespace, casing, and key ordering; sensitive to model swaps, tool list, and retrieval context. Python port of @mukundakatta/semantic-cache-key.
...moreLint Claude Code SKILL.md files for frontmatter, required fields, descriptions, and hardcoded secrets. Stdlib-only Python port of @mukundakatta/skillint.
...moreMukunda Katta
Token-aware text chunker for RAG ingestion. Sentence-respecting, overlap-friendly.
Parse partial / truncated / messy JSON for LLM tool calls and structured outputs.
Detect system prompt leakage in LLM model outputs via known patterns, configured-prompt substring matching, and unique fingerprint phrases. Python port of @mukundakatta/system-prompt-leak-scan.
...moreThread-safe shared token + USD budget for concurrent LLM tasks. Raises BudgetExceeded on push past cap. Reserve/commit two-phase API for fan-out workloads. Zero runtime deps.
...moreValidate LLM tool-call payloads with small JSON-like contracts. Python port of @mukundakatta/tool-call-contracts.
Truncate LLM tool output with head/tail/middle/middle_lines strategies. UTF-8 safe, zero runtime deps.
Score (query, document) pairs for vector/RAG poisoning signals: vector-text mismatch, instruction-like payloads, NaN, suspiciously round numbers. Python port of @mukundakatta/vector-poison-score.
...moremukundakatta
Validate LLM tool-call payloads with small JSON-like contracts.
mukundakatta
Scaffold and validate Codex skills from the command line.
mukundakatta
Prompt-injection risk scanner for untrusted AI context.
mukundakatta
Plan model fallback chains from capability, cost, and health data.
mukundakatta
Policy-check agent tool calls before execution.
mukundakatta
Heuristic quality metrics for RAG retrieval and grounded answers.
mukundakatta
Find stale RAG chunks by age, version, and freshness requirements.
mukundakatta
Enforce document ACLs after retrieval and before prompting.
mukundakatta
Pack context chunks into a budget by relevance and priority.
mukundakatta
Streaming JSON parser that yields partial valid trees as tokens arrive. Built for LLM tool-call payloads, structured output streams, and any place a regular JSON.parse waits too long.
...moremukundakatta
MCP server: snapshot tests for tool-call traces. Capture, normalize, and diff agent tool-use traces to catch silent regressions. Wraps @mukundakatta/agentsnap for Claude Desktop, Cursor, Cline, Windsurf, and Zed.
...moremukundakatta
Fit your messages into the LLM context window. Token-aware truncation with multiple strategies (drop-oldest, drop-middle, priority), pluggable tokenizers, zero dependencies.
...moremukundakatta
Detect system prompt leakage in model outputs.
mukundakatta
MCP server for agentvet: validate LLM-generated tool-call args before execution. Returns a typed error with an LLM-friendly retry hint when the args don't match.
...moremukundakatta
MCP server exposing prompt-injection-shield: scan untrusted text for prompt-injection signals, score risk, strip dangerous lines. Built for Claude Desktop, Cursor, Cline, Windsurf, and Zed.
...moremukundakatta
MCP server: AWS Bedrock model capability lookup + feature pre-check + PII-safe response redaction. Wraps the Python bedrock-ops library.
...moremukundakatta
Score retrieved documents for vector/RAG poisoning signals.
mukundakatta
MCP server: decode JWT to inspect header and claims (no verification).
mukundakatta
MCP server: encode/decode base64, base64url, hex, and percent-encoding.
mukundakatta
MCP server: encode and decode HTML entities, URL-encoding, and Unicode escapes.
mukundakatta
MCP server: parse and serialize .env files. Handles quotes, comments, multiline.
mukundakatta
MCP server: hash files or strings with sha-2, sha-3, blake2/3, or xxhash.
Mukunda Katta <[email protected]>
Build and validate lightweight AI model / data / tool manifests. Python port of @mukundakatta/ai-supply-chain-manifest.
Mukunda Rao Katta <[email protected]>
Small local jailbreak and prompt-injection fixture set for tests. Python port of @mukundakatta/jailbreak-corpus-mini.
Mukunda Katta <[email protected]>
Plan model fallback chains from capability, cost, and health data. Python port of @mukundakatta/model-fallback-planner.
Mukunda Katta <[email protected]>
Policy-based model routing by capability, cost, latency, and privacy. Python port of @mukundakatta/model-router-policy.
Mukunda Rao Katta
CLI wrapper for prompt-injection-shield-py: scan a file or stdin for prompt-injection patterns.
Mukunda Rao Katta <[email protected]>
Track untrusted tool output before it enters prompts or actions. Python port of @mukundakatta/tool-result-taint.