QG016 Unmodelled path

Type	Vulnerability
Severity	by action, never below High
Front end	an agent graph, and a Colang policy where its actions map to the catalogue's tools
Guardrail category (Vassilev 2026)	E. Action & Execution Guardrails
AI RMF functions (NIST AI 100-1)	GOVERN, MANAGE
OWASP	ASI10:2026 Rogue Agents
Gates by default	yes, always

Raised when

An open attacker campaign realised a privileged call on a graph where the engine found no way through.

Why it matters

The engine verified the graph it was given and an attacker found a way through the graph that runs: the export and the running graph disagree, or the agent acts outside the policy the engine verified. Nothing else this product reports is more urgent.

How to fix it

Re-export the graph from the code that runs and re-scan; then compare the attacker's trace with the export to find what the export is missing.

OWASP entries

LLM01:2026 Prompt Injection, ASI10 Rogue Agents, from the OWASP Top 10 for LLM Applications 2026 and for Agentic Applications 2026. The mapping holds because this is the agent acting outside the policy the engine verified, found by injection. A finding under a rule mapped to LLM03:2026 carries the root cause in brackets as owasp_root_cause.

What the finding carries

A location at the node's source -- the file and line of its callable, or the line of agent.yaml that declared it -- falling back to the export's JSON Pointer; a fingerprint over the rule, the tool, the control, the node and a digest of the node's declaration, so moving unrelated nodes keeps it and changing the node's tools or edges makes a new one; the witness as a path through the graph; and a structural fix. A graph finding is replayed in the team's environment, with every tool stubbed and a scripted model making the witness's choices; until then it reads not yet replayed. See the catalogue for the other rules.