# Agent trajectories as programs: fingerprinting and programming coding-agent behavior

DevFeed: [Agent trajectories as programs: fingerprinting and programming coding-agent behavior](<https://devfeed.tech/articles/agent-trajectories-as-programs-fingerprinting-and-programming-coding-agent-behavior-116167.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2606.16988>)

Author: Hamidah Oderinwale

Published: 2026-11-04T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>), [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [agentic-coding](<https://devfeed.tech/topics/agentic-coding.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [agent observability](<https://devfeed.tech/topics/agent-observability.md>)

Tags: [about](<https://devfeed.tech/tags/about.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [apply](<https://devfeed.tech/tags/apply.md>), [approach](<https://devfeed.tech/tags/approach.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [auditing](<https://devfeed.tech/tags/auditing.md>), [avoid](<https://devfeed.tech/tags/avoid.md>), [behavior](<https://devfeed.tech/tags/behavior.md>), [bench](<https://devfeed.tech/tags/bench.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [coding](<https://devfeed.tech/tags/coding.md>), [coding-agent](<https://devfeed.tech/tags/coding-agent.md>), [compare](<https://devfeed.tech/tags/compare.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [define](<https://devfeed.tech/tags/define.md>), [develop](<https://devfeed.tech/tags/develop.md>), [dimensions](<https://devfeed.tech/tags/dimensions.md>), [programming](<https://devfeed.tech/tags/programming.md>)

## AI overview

ProcGrep analyzes agent traces to compare how agents solve tasks, beyond whether they succeed. It uses compressed procedural representations built through emergent vocabulary induction to capture recurring behavioral patterns while reducing surface variation. In a comparison of ten agents, fingerprints identified the agent behind an unseen trajectory with 85.7% accuracy. On SWE-Bench, trajectories were most similar for models from similar release periods and for related models such as a distilled student and its teacher; their Jensen-Shannon divergence was 0.25, about half that of other model pairs. The work presents ProcGrep as a library for top-down procedural auditing and evaluation. These findings are limited to the reported agents and dataset.

## Source excerpt

arXiv:2606.16988v2 Announce Type: replace Abstract: Benchmark scores tell you what an agent got right; they do not tell you how it got there. In this work, we introduce methods for comparing agents procedurally in different contexts, where the model, tasks, and approaches vary. We compare ten agents and find that they are identifiable by their behavioral habits, which we define as fingerprints: a probe over these procedural signatures attributes an unseen trajectory to the correct agent at 85.7% accuracy. We develop procedural representations for agent problem-solving procedures with an emergent vocabulary induction technique which is meant to be maximally compressive to avoid surface-level variation while being expressive enough to unveil the quirks of the models' patterns. We apply our framework to the software engineering evaluation dataset SWE-Bench to study the structural distinctness of agent trajectories and find that behavior is most similar between models from similar release periods and those that are distilled from one another (e.g., a distilled student model and its teacher have a Jensen-Shannon divergence of 0.25, about half the distance between other model pairs). As more models saturate evaluations, we believe that it will be important to probe model behavior along more holistic dimensions than success rates alone. We introduce ProcGrep, a library for auditing and evaluating agents for how they approach tasks at a procedural level given their traces in a top-down fashion. ProcGrep is available at https://github.com/hamidahoderinwale/procgrep.