DEVELOPER NEWS STREAM
Direct logs, engine updates, and framework notifications parsed from curated RSS feeds and announcements, updated hourly.
Mukunda Rao Kattaclaude-stream-rs: parse Anthropic SSE streaming responses into a typed Rust enum
A zero-HTTP, zero-async-runtime SSE parser that turns Anthropic streaming bytes into a typed Event enum. Feed it bytes from any source, get events out.
Mukunda Rao KattaI run 30 agents in parallel. They share one budget. Here is the pool primitive that makes that safe.
A small Rust crate for thread-safe shared token and USD budgets across concurrent agents. Reserve plus commit, BudgetExceeded fail-fast, zero allocations on the hot path.
Mukunda Rao Kattatool-loop-guard-rs: Break Agent Loops Before They Drain Your Budget
A Rust crate that detects repeated identical tool calls within a sliding window. Stops runaway agent loops before they hit your API invoice.
Mukunda Rao Kattallm-retry: Full-Jitter Exponential Backoff for LLM API Calls in Rust
A runtime-agnostic Rust crate with full-jitter exponential backoff, built-in retryable status code lists for Anthropic, OpenAI, Bedrock, and Gemini, and both sync and tokio async support.
Mukunda Rao Kattallm-budget-window: Per-Minute and Per-Hour Token Caps That Actually Work
A Rust crate with sliding time windows for token and USD budgets. Blocks before the expensive call, not after.
Mukunda Rao Kattatool-arg-fuzzy-rs: Fuzzy Match LLM Enum Args to Valid Values in Rust
A Rust crate that cascades exact, case-insensitive, prefix, and substring matches to recover from LLM enum hallucinations without Levenshtein.
Mukunda Rao Kattaagentvet-rs: Validate Tool Args Before Your Agent Executes Them (Rust)
A Rust crate that wraps tool functions with JSON Schema arg validation and throws a retry-friendly error before any real work happens.
Mukunda Rao KattaRust: Tag Your Agent Tools With Side-Effect Metadata Before Dispatching Them in Parallel
tool-side-effects-tag-rs gives every tool a HashSet of READ/WRITE/IDEMPOTENT/DESTRUCTIVE tags so your dispatcher can check safety before firing concurrent calls.
Mukunda Rao Kattaagentsnap-rs: Snapshot Testing for Rust Agent Tool-Call Traces
Record your Rust agent's tool-call trace once, then fail fast whenever the calls, args, or order change unexpectedly.
Mukunda Rao Kattaagenttrace-rs: Group LLM Calls into Named Runs and Get Cost Breakdowns in Rust
A Rust crate that records LLM calls into named runs and reports total cost, p50/p95 latency, and per-model breakdowns without any external service.
Mukunda Rao Kattaagentguard-rs: Domain Allowlists for Rust AI Agent Network Calls
Enforce a declarative domain allowlist on every HTTP/HTTPS call your Rust agent makes, so prompt injection cannot exfiltrate data to an attacker-controlled host.
Mukunda Rao KattaStop Hammering a Broken API: Circuit Breakers for LLM Calls in Rust
When your retry logic becomes the disaster. A thread-safe Closed/Open/HalfOpen circuit breaker for LLM API calls, with zero dependencies.
hiyoyoBuilding a Card Game AI with Reinforcement Learning — Implementation Details#2
Introduction This is a follow-up to my previous article. If you haven't read it yet, check...
Bruno BorgesTOML Schema is live
TOML Schema is now available for testing and evaluation, and now has a home:...
pretty ncubeWhen Server Growth Torpedoes Your Treasure Hunt Engine
The Problem We Were Actually Solving I was tasked with scaling our treasure hunt engine to...
pretty ncubeThe Moment the Jaeger Tracer Exhausted Itself and What We Switched To
The Problem We Were Actually Solving Our treasure-hunt engine at Veltrix was not exploding; it was...