Everything to run an AI agent fleet

See what your agents do, cap what they spend, and keep an audit trail.

Curated by Code Recycle Editorial

1. 1

   @amos-ui/agent-timelineComponent✓ Free

   Every AI product rebuilds the same layer: what is the agent doing right now, what did that tool call cost, who approves the risky step. The chat frameworks render messages and leave that part to you.

   Why it's here: See what the agent is doing: tool calls, reasoning, cost, approval gates.

2. 2

   True Token LedgerComponent✓ from $9

   Your cost dashboard multiplies tokens by a rate. Your invoice charges for billing events. On a cache-heavy workload those are different numbers.

   Why it's here: Your cost dashboard multiplies tokens by a rate. Your invoice charges for billing events.

3. 3

   Spend Ceiling LimiterComponent✓ from $19

   A rate limiter whose rejection path was the memory leak: 500 callers left 500 entries for 40 served requests.

   Why it's here: Hard spend limits — from the limiter whose own rejection path was the memory leak.

4. 4

   Lane GuardComponent✓ from $29

   Two agents, one checkout: agent B's commit silently contains agent A's unfinished work.

   Why it's here: Two agents, one checkout: agent B's commit silently contains agent A's unfinished work.

5. 5

   Tool-Schema Drift DetectorComponent✓ from $9

   Nothing connects a tool schema to the function it calls. Rename a parameter and it breaks on the next model call, in production, days later.

   Why it's here: Nothing connects a tool schema to the function it calls. Rename a parameter and it breaks.

6. 6

   Tool-Result Size GuardComponent✓ from $9

   Truncation happens at a byte offset, never a structural one. A JSON payload cut mid-string parses into something silently incomplete.

   Why it's here: Truncation happens at a byte offset, never a structural one.

7. 7

   Tamper-Evident Hash-Chained Audit LogComponent✓ from $19

   Hash each entry independently and verification passes while entries are deleted, reordered or truncated.

   Why it's here: An audit trail that actually detects deletion, reordering and truncation.

8. 8

   THE PROMPT BREAKERgame✓ Free

   Seven levels of prompt injection against a hardened agent. Every defense is real. Every one is partial.

   Why it's here: Seven levels of prompt injection against a hardened agent. Every defense is partial.

9. 9

   langflowApplicationUnclaimed · identity unverifiedFree

   Langflow is a powerful tool for building and deploying AI-powered agents and workflows.

   Why it's here: Free visual builder if you're still designing the agents themselves.