Agent loop with no step limit
MediumAV-RUNTIME-014Runtime safetyCore pack
What it finds
The agent's reasoning/tool loop is configured with no maximum number of steps.
Why it matters
A confused or manipulated agent can loop forever — burning tokens and API budget, hammering downstream systems, or repeating an action many times.
What an attacker could do: A single bad task could run up unbounded model cost or repeat a side-effecting tool call indefinitely.
How to fix it
Set an explicit step/iteration limit (typically 10–30), a wall-clock timeout, and a per-task token budget.
Pro shows the corrected code for your exact file, step by step, and re-scans to confirm the fix. See plans
Check details
- Finding code
- UNBOUNDED_AGENT_LOOP
- How it detects
- Looks for a specific risky pattern in the code or config.
- Confidence
- Definite: the risky setting is written plainly in the file.
- Scans
- Any text-based agent source or configuration (JS/TS/Python/JSON/YAML)
- OWASP LLM Top 10
- LLM04 - Model Denial of Service, LLM08 - Excessive Agency
- NIST AI RMF
- MG-2.2
- SOC 2
- A1.1, CC7.2
Check your agent for this
Paste your agent code or MCP config. The free scan runs all 124 checks in about 10 seconds — no account needed.
Related checks
- Command injection riskAV-RUNTIME-004 · Critical
- Unsafe dynamic code executionAV-RUNTIME-005 · Critical
- Privileged or host-mounted container configurationAV-RUNTIME-007 · Critical
- SQL query built from unescaped inputAV-RUNTIME-009 · Critical
- Unsafe deserialization of untrusted dataAV-RUNTIME-010 · Critical
- Prompt injection surface detectedAV-RUNTIME-002 · High