Popular repositories Loading

1. claude-context-audit claude-context-audit Public

   Audit what your CLAUDE.md actually costs: token weight per file/import/skill, duplicate directives, and a trim list. Offline, report-only, zero deps.

   Python 2

2. claude-cost-attribution claude-cost-attribution Public

   Cache-aware, per-skill Claude Code cost attribution CLI. Reads local transcripts, attributes token cost per project/skill/subagent, plus a plan-cap burn-rate gauge. Stdlib only, $0, no network.

   Python 1

3. llm-judge-calibration llm-judge-calibration Public

   Calibrate an LLM-as-judge against human labels: Cohen's kappa, accuracy CI, position and verbosity bias probes, and an honest reliable-for-X-not-Y verdict. Stdlib only, $0 offline path.

   Python 1

4. rag-chunking-eval rag-chunking-eval Public

   Measure RAG retrieval quality across chunking configs (size x overlap x splitter): hit-rate@k and MRR per config, so you pick with data not vibes. Hand-rolled BM25, stdlib only, offline.

   Python 1

5. n8n-retry-runbook-reference n8n-retry-runbook-reference Public

   The automation reliability layer, shown not claimed: an n8n workflow with a visible retry/idempotency/dead-letter branch, a runbook, a monitoring pattern, and a $0 offline simulator that proves it.

   Python 1

6. secret-scrubber secret-scrubber Public

   Find and redact secrets, API keys, and PII in a log or prompt before you paste it into an AI. Runs entirely in your browser, nothing is uploaded.

   JavaScript 1