pj bee430cf72 fix(runner): no silent generator fallback, and llm on web ...

--generator llm against a spec declaring no generator = llm(...) logged a
warning and ran the seeded picker. For a comparison campaign that is silent
arm corruption: the run completes, the directory looks correct, and the wrong
policy drove it. It is now fatal.

pickSources also returned the V8 source for both action and extractor on web
before it looked at the generator, so the llm policy was unreachable there.
The two axes are now independent: the driver picks the extractor source, the
flag picks the action source, and llmSource composes with either because the
runner populates the candidate list and screenshot on every platform.

Claude-Session: https://claude.ai/code/session_01A5KmftdEJ49A9z5mF5ESrX

2026-08-12 18:56:08 +05:30

..

_oracle

iOS simulator driver: Go-native companion-backed backend (#62)

2026-06-08 19:10:54 +05:30

android

WIP: Drive physical Android devices over USB (#67)

2026-06-11 10:10:05 +05:30

bundler

llm action-selection backend (#68)

2026-07-31 21:12:00 +05:30

driver

fix ltl semantics and unify action enumeration (#71)

2026-08-12 18:06:04 +05:30

hierarchy

llm action-selection backend (#68)

2026-07-31 21:12:00 +05:30

ios

Physical-device iOS support (#64) (#66)

2026-06-09 18:38:52 +05:30

llmclient

llm action-selection backend (#68)

2026-07-31 21:12:00 +05:30

ltl

fix ltl semantics and unify action enumeration (#71)

2026-08-12 18:06:04 +05:30

replay

iOS simulator driver: Go-native companion-backed backend (#62)

2026-06-08 19:10:54 +05:30

runner

fix(runner): no silent generator fallback, and llm on web

2026-08-12 18:56:08 +05:30

sidecarassets

test: full test-suite refactor sweep (#61)

2026-06-06 13:59:08 +05:30

testrun

feat(cli): add --arm and populate run meta from it

2026-08-12 18:14:07 +05:30

trace

feat(trace): record arm membership and host in meta.json

2026-08-12 18:12:28 +05:30

verifier

fix ltl semantics and unify action enumeration (#71)

2026-08-12 18:06:04 +05:30