Repository navigation

feat: agentic wf testing in harness worker (+ fix agent-e2e) - #534

Open

chrishagglund-ship-it wants to merge 18 commits into

mainfrom

harness-agentic-testing

Open

feat: agentic wf testing in harness worker (+ fix agent-e2e)#534
chrishagglund-ship-it wants to merge 18 commits into
mainfrom
harness-agentic-testing

Commits on Nov 9, 2026
- feat(harness): experimental agentic testing proof of concept (untested)
  
  chrishagglund-ship-it
  and
  claude
  committed
  a04ddfb
- feat(harness): execution links, --open and --verbose for try_agentic.py
  
  chrishagglund-ship-it
  and
  claude
  committed
  0b89103
- start agent in a way that spawns stop_when worker
  chrishagglund-ship-it
  committed
  55944b6
- feat(harness): --stop-same-turn and a report built from the execution
  
  chrishagglund-ship-it
  and
  claude
  committed
  ded4ecb
- feat(harness): rewire stop_when in harness runs by default
  
  chrishagglund-ship-it
  and
  claude
  committed
  21f2362
- feat(harness): report each agentic run's outcome and story
  
  chrishagglund-ship-it
  and
  claude
  committed
  e4ba43c
- feat(harness): single-replica agentic deployment for certification-aws
  
  chrishagglund-ship-it
  and
  claude
  committed
  3037c34
- feat(harness): end agentic runs the harness can't finish correctly
  
  chrishagglund-ship-it
  and
  claude
  committed
  37a6c50
- ci(agent-e2e): cap pytest below 9.2 to keep xdist groups working
  
  chrishagglund-ship-it
  and
  claude
  committed
  2f6f131
Commits on Nov 10, 2026
- helper script to walk through wf and show how story evolved until wf finished
  chrishagglund-ship-it
  committed
  9d9813c
- ci(agent-e2e): fail fast with a named cause when the LLM path is broken
  
  chrishagglund-ship-it
  and
  claude
  committed
  b472fa5
Commits on Nov 11, 2026
- ci(agent-e2e): check LLM providers first, drop the canary agent run
  
  chrishagglund-ship-it
  and
  claude
  committed
  888db13
- ci(agent-e2e): run the suites on gpt-6-luna
  
  chrishagglund-ship-it
  and
  claude
  committed
  8a664e2
- TEMP: run test_image_openai to confirm its OpenAI error
  
  chrishagglund-ship-it
  and
  claude
  committed
  10968d8
- e2e: skip test_image_openai, dall-e-3 is gone from OpenAI's API
  
  chrishagglund-ship-it
  and
  claude
  committed
  5bcc918
- TEMP: run test_audio_openai to see its OpenAI error
  
  chrishagglund-ship-it
  and
  claude
  committed
  71335b7
- e2e: skip test_audio_openai, the CI OpenAI project can't use tts-1
  
  chrishagglund-ship-it
  and
  claude
  committed
  9d7fc73
- e2e: note the image test's gpt-image-* models aren't enabled in CI
  
  chrishagglund-ship-it
  and
  claude
  committed
  af3940e

Commits

Commits on Nov 9, 2026

Commits on Nov 10, 2026

Commits on Nov 11, 2026