loading…
Search for a command to run...
loading…
Description: Classify AI/API execution failures. * Returns WAIT, CAP, or STOP with * corpus-grounded precedent from 63 * confirmed production failures.
Description: Classify AI/API execution failures. * Returns WAIT, CAP, or STOP with * corpus-grounded precedent from 63 * confirmed production failures.
The reliability layer for AI agents.
When your agent fails at an execution boundary, Pitstop tells it what happened, what it costs, and how to recover.
Agents pull in when execution fails. They get what they need and get back on the road. Every stop makes the next stop faster.
POST a 429 status + headers, get back a classification:
curl -s -X POST https://web-production-273d3.up.railway.app/classify \
-H "Content-Type: application/json" \
-d '{"status":429,"headers":{"retry-after":"30"},"provider":"anthropic"}'
Returns: WAIT / CAP / STOP + fix shape + matched corpus precedent.
63 production failure receipts. 14 confirmed and merged by maintainers.
github.com/SirBrenton/pitstop-truth
pitstop-truth (corpus)
↓
ChromaDB (local vector index)
↓
/classify (Railway endpoint)
↓
MCP server (agents call this)
[email protected] pitstop.dev @SirBrenton on GitHub
Run in your terminal:
claude mcp add pitstop -- npx