Tigerless_ailab

4 ms·

1. 1.▲
   Show HN: Autoharness – a self-learning, maintaining skill layer for Claude Code(github.com)

   3 points by Tigerless_ailab 3mo ago|0 comments
2. 2.▲
   Ask HN: How do you measure whether your coding agent follows its rules?

   2 points by Tigerless_ailab 3mo ago|0 comments
3. 3.▲
   by Tigerless_ailab 4mo ago

   Yes — that's exactly why I built it. Log-based tools only account for roughly half of the token cost I wanted to see where the other ~50% goes: system prompts, tool schemas, MCP overhead, plus per-tool stats on tool use and tool outp
4. 4.▲
   by Tigerless_ailab 4mo ago

   Not really. Codeburn tells you how much a chat/code cost, we tell you what's inside your request from API: how much does system prompts, schemas, MCP blocks, thinking, cache reads/writes, tool calls, and outputs cost.
5. 5.▲
   by Tigerless_ailab 4mo ago

   Lol, actually not, it's mine.
6. 6.▲
   by Tigerless_ailab 4mo ago

   Yes! Install it and you get a live view of where every session's tokens come from — system prompt, each tool's inputs and outputs, the tool schemas, etc.
7. 7.▲
   I stopped reading logs but the raw API requests – here's where my token go(github.com)

   4 points by Tigerless_ailab 4mo ago|2 comments