[BUG] / [FEATURE] : Preserve generated output on reached Limits issues / Cost Issues #261
Description
Describe the bug
I run strix on github workflow with a paid claude subscription and run into a rate limit. Output for round about 2$ was wasted, as it has been thrown away.
To Reproduce
Steps to reproduce the behavior:
- Set up Strix as Github Workflow
- Set up Claude Paid Subscription API Key
- Run strix on an (not so large) codebase
- With bad luck the procedure was too intensive
❌ Penetration test failed: LLM request failed: Rate limit exceeded
╭────────────────────── 🔍 Live Penetration Test Status ───────────────────────╮
│ │
│ 🦉 Running penetration test... │
│ │
│ 🧠 Model: *** │
│ 🔍 Vulnerabilities: 0 │
│ 🤖 Agents: 7 • 🛠️ Tools: 87 │
│ 📥 Input: 2.1M • ⚡ Cached: 1.6M │
│ 📤 Output: 23.7K • 💰 Cost: $2.6737 │
│ │
╰──────────────────────────────────────────────────────────────────────────────╯
Error: Process completed with exit code 1.
Reference: https://github.com/mrwiora/tpm2-kira/actions/runs/20994975512/job/60351631624
Expected behavior
- best outcome would be a finished resultset
- when it aborts the intermediate output that has been generated should be preserved somewhere
Screenshots
System Information:
- Claude API Key with spending limit on 20$
- Github Workflow https://github.com/mrwiora/tpm2-kira/blob/copilot/add-ai-pentest-scan-workflow/.github/workflows/strix-security-scan.yml
Additional context
I answered to a similar bug, but obviously nobody gets notified there anymore, as this one is closed.
#80 (comment)
I requested support from Claude, but this seems to be not of interest for them.