> Markdown version of [/events/conference-india-2026/sessions/2083-objection-put-your](https://www.wearedevelopers.com/events/conference-india-2026/sessions/2083-objection-put-your). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Objection! Put Your Agent’s Code on Trial, Where Sessions Testify and Checkpoints Are Evidence


- **Event:** Conference India 2026

## Description

Objection! Put Your Agent’s Code on Trial, Where Sessions Testify, Checkpoints Are Evidence, and Pull Requests Are Hearsay

Description: Every pull request asks you to rule on hearsay. The diff says what changed, but the only witness who knows why, your agent’s session, never takes the stand.

That shows up in three ways:
- Reviewers rule without the testimony. The prompt, the plan, and the approaches the agent dropped stay buried in a session nobody brings to court, so approvals come down to a shrug and an “LGTM.”
- The evidence disappears after the verdict. Once code merges, nothing ties a strange line back to the session that wrote it, so nobody can tell a bug from a deliberate decision.
- Every case starts without precedent. Past rulings never reach the next agent, so it repeats mistakes review already caught and burns tokens relearning the codebase.

In this hands-on workshop, your agent’s code goes on trial, and nobody opens a pull request. Entire records each session as the agent’s testimony, and each commit seals a checkpoint that enters the code into evidence along with the prompts and decisions behind it. Teams take turns as prosecution and defense. One side traces suspicious lines back to the exact prompt that created them, while the other rebuilds the agent’s reasoning from its session to argue for the merge.

You’ll learn how to present each case on a Trail with its sessions attached and call in review agents as expert witnesses that score each push. You’ll let gates deliver the verdict, then turn each ruling into precedent with Entire Brain, so the next agent walks into court already knowing the case law. You’ll finish with a shareable rap sheet for your agent: every charge, every acquittal, and the precedents it set.

## Speaker

### [Abhi Singh](https://www.wearedevelopers.com/@abhi-singh)

Solutions Engineer at Entire


## Related talks at this congress

- [The Developer’s Guide to the Agentic Git Forge: Agents That Remember](https://www.wearedevelopers.com/events/conference-india-2026/sessions/2077-the-developer-s) — Haimantika Mitra
- [The Developer’s Guide to the Agentic Git Forge: Merge at Agent Speed](https://www.wearedevelopers.com/events/conference-india-2026/sessions/2078-the-developer-s) — Haimantika Mitra
- [More Agents. Fewer Tabs. Meet the GitHub Copilot App](https://www.wearedevelopers.com/events/conference-india-2026/sessions/2032-more-agents-fewer) — Romaan Shair
- [Did the Model Improve, or Just the Score? A Developer’s Guide to LLM Evaluation](https://www.wearedevelopers.com/events/conference-india-2026/sessions/2012-did-the-model) — Jack Fitzsimons

## Watch remotely

Can’t make it to Bengaluru? Watch this session live with Pro. You also get:

- All full videos, bookmarks, and playlists
- World Congress livestreams

[See pricing](https://www.wearedevelopers.com/pricing)

## Links

- [Get tickets](https://www.wearedevelopers.com/conference-india/tickets)