AgentClash

Use case

Evaluate research agents with evidence quality

Research agents live or die on sourcing, synthesis, and artifact quality. AgentClash scores whether an agent found the right evidence, cited it correctly, and finished the investigation under budget.

live eval
gate: pass

Candidate

92correct patch, low cost

Baseline

88stable reference run

Control

73missed edge case

replay timeline

1loaded task inputs and tool policy
2ran sandbox actions and captured artifacts
3scored trajectory and validator evidence
4attached scorecard and release verdict

ci verdict

Candidate clears release gate

Correctness improved, latency within budget, and required artifacts were preserved for review.

agentclash run create --follow

Research eval signals

Built for reviewable agent decisions

Measure coverage, citation quality, artifact completeness, and whether the agent stopped with a useful deliverable.

OpenTelemetry-compatible trace import

Pinned datasets and golden test cases

Baseline versus candidate regression checks

Replay trails for tool calls, outputs, and artifacts

Scorecards for correctness, cost, latency, and evidence

CI gates for prompt, model, RAG, and tool changes

Workflow

Research eval workflow

Import the evidence

Start from OpenTelemetry traces, curated datasets, support transcripts, or a real failure your team already saw.

Pin the baseline

Record the current accepted behavior so every prompt, model, RAG, or tool change has a fair comparison point.

Replay the evidence

Inspect tool calls, outputs, artifacts, latency, cost, and judge evidence when a candidate gets worse.

Gate the release

Compare candidate and baseline runs, then fail CI before a regression reaches users.

FAQ

Research agent evaluation FAQ

What should research agent evaluation score?

Task completion, evidence quality, artifact completeness, tool discipline, and whether the final synthesis matches the sources collected.

Can evals include web or file tools?

Yes. Challenge packs define which tools agents may use and which artifacts must be produced for a pass.

How do teams debug a failed research run?

Replay shows each search, fetch, note, and synthesis step so reviewers can see where the investigation went off track.