Search results

tool-call-f1 misses OpenAI Responses function_call outputs

Status: Open.

#10649 In promptfoo/promptfoo;

· JuSe123456 opened on Sep 5, 2026

test(examples): cover trajectory step status in OpenTelemetry agent example

Status: Open.

#10609 In promptfoo/promptfoo;

· Asthenia0412 opened on Sep 2, 2026

feat(assertions): evaluate trajectory step execution status

Status: Open.

#10607 In promptfoo/promptfoo;

· Asthenia0412 opened on Sep 2, 2026

Should avgLatencyMs reflect replayed latency for cache-hit responses?

Status: Open.

#10595 In promptfoo/promptfoo;

· roy-tong opened on Sep 1, 2026

Model-graded assertions fall back to defaultTest.provider — the provider under test — when no grader is set, and the grader-override docs don't mention it

Status: Open.

#10581 In promptfoo/promptfoo;

· mrhpython opened on Sep 1, 2026

agent-rubric: the only prompt-injection control I found in a 9-framework survey

Status: Open.

#10575 In promptfoo/promptfoo;

· idilgozel opened on Aug 31, 2026

Discussion: Deeper CI/CD PR integration and local Ollama command execution?

Status: Open.

#10548 In promptfoo/promptfoo;

· mukundzha opened on Aug 29, 2026

openai:chat apiBaseUrl must include /v1; default graders still use gpt-5.5

Status: Open.

#10526 In promptfoo/promptfoo;

· cursor opened on Aug 28, 2026

opencode:sdk provider never threads per-call traceparent to the OpenCode server, breaking trajectory:* assertions

Status: Open.

#10518 In promptfoo/promptfoo;

· b-per opened on Aug 27, 2026

Question: stable per-assertion output for an external adapter

Status: Open.

#10514 In promptfoo/promptfoo;

· GautamTalksDev opened on Aug 26, 2026

Proposal: opt-in live pricing fallback via OpenRouter's public models API for unrecognized models

Status: Open.

#10513 In promptfoo/promptfoo;

· kimyongjun0129 opened on Aug 26, 2026

Benchmarked promptfoo against two third-party targets, plus a question about plugin selection

Status: Open.

#10505 In promptfoo/promptfoo;

· qatration opened on Aug 26, 2026