Commit Graph
59 Commits
Author SHA1 Message Date
ciregenz cdbdec775b [eric] browser: merge-verify, a confirmed send is the proof, no separate verify turn + trailing-read telemetry 2026-06-06 04:35:03 -07:00
ciregenz 6dc00e40b5 [eric] browser: think-shorter, no redundant prose beside action tools + output telemetry 2026-06-06 04:14:46 -07:00
ciregenz 852968d7b2 [eric] browser: candidate scan also covers the pre-nav landing, shared helper 2026-06-06 00:21:31 -07:00
ciregenz 56eea88153 [eric] browser: wire route hints, entry pre-nav, auto candidate scan and llm_ms timing into the loop 2026-06-06 00:17:05 -07:00
ciregenz 92df6ea4f0 [eric] browser: send-probe gate catches focus-type payloads, batched Enter blocked while composer pending 2026-06-05 23:30:23 -07:00
ciregenz fcc02ce77a [eric] browser: evaluation_previous optional, output tokens drive turn latency and the field restates attached state 2026-06-05 22:23:18 -07:00
ciregenz 042e7db7d5 [eric] browser: pruning-safe incremental cache marker, transcript finally cache-hits past the system prompt 2026-06-05 22:22:06 -07:00
ciregenz 339d7563b9 [eric] browser: hard send-guard on live batches, irreversible steps physically cannot ride along 2026-06-05 22:19:55 -07:00
ciregenz 9af337585b [eric] browser: re-land schema-forced batching + redundant-read nudge (zombie forensics exonerated both) 2026-06-05 22:16:40 -07:00
ciregenz e7099cc888 Revert "[eric] browser: schema-forced batching, solo mutators hidden, ClickIndex reserved for sends and fills"
This reverts commit 2fe8e0f62f.
2026-06-05 20:07:08 -07:00
ciregenz 065c006a52 Revert "[eric] browser: nudge redundant solo re-reads, the biggest turn sink after batching"
This reverts commit 8c060da74b.
2026-06-05 20:07:08 -07:00
ciregenz dcba6d4e35 Revert "[eric] browser: redundant-read flag survives intervening reads, catches re-list clusters"
This reverts commit 1a86ee2ad0.
2026-06-05 20:07:08 -07:00
ciregenz 1a86ee2ad0 [eric] browser: redundant-read flag survives intervening reads, catches re-list clusters 2026-06-05 19:51:30 -07:00
ciregenz 8c060da74b [eric] browser: nudge redundant solo re-reads, the biggest turn sink after batching 2026-06-05 19:46:22 -07:00
ciregenz 2a3e5315ef [eric] browser: park send-skill prefix replay behind a flag until settle and detour fixes land 2026-06-05 18:22:14 -07:00
ciregenz 69f15ce9ad [eric] browser: batched click_index distills via per-sub identities, batched runs record skills again 2026-06-05 18:07:32 -07:00
ciregenz 2fe8e0f62f [eric] browser: schema-forced batching, solo mutators hidden, ClickIndex reserved for sends and fills 2026-06-05 17:56:17 -07:00
ciregenz 1480fddf51 [eric] browser: replay clicks scroll-and-retry when off-screen, long names resolve by stable prefix 2026-06-05 17:21:50 -07:00
ciregenz 7a30cc97cd [eric] browser: prefix replay also fires at the deferred host re-check, miss paths logged 2026-06-05 16:30:41 -07:00
ciregenz 88ba59f3b4 [eric] browser: send-skill prefix replay hands off at the composer, under-batching nudges 2026-06-05 16:12:45 -07:00
ciregenz 2c55d17f50 [eric] browser: dead-dashboard dispatches fail in 1s instead of burning turns, recovery skips systemic causes 2026-06-05 15:56:38 -07:00
ciregenz e37ba6c66b [eric] browser: skill recording judges the task ask, OUTCOME boilerplate no longer blocks learning 2026-06-05 13:09:02 -07:00
ciregenz 99bd15fdb1 [eric] browser: attached page state shrinks to changed rows, unchanged indices stay valid 2026-06-05 11:14:13 -07:00
ciregenz 059185dd4b [eric] browser: kill the double-fill, textbox rows show live value, typing echoes verified read-back, expect demands literal page text 2026-06-05 10:29:12 -07:00
ciregenz 20e1558d70 [eric] browser: auto-settle mutating actions and attach fresh page state to results, act and see in one turn 2026-06-05 09:35:51 -07:00
ciregenz 17eac6412d [eric] browser: derive the reuse host from the task text when the dispatch has no url param 2026-06-05 01:43:09 -07:00
ciregenz 65427f33a0 [eric] browser: reuse the parent's same-host card on re-dispatch, stacked webviews wedge each other 2026-06-05 01:38:58 -07:00
ciregenz 88088a9564 [eric] browser: BrowserExtract, aux-model schema extract + prompt docs for ctx and stable indices 2026-06-05 00:55:14 -07:00
ciregenz 262d48ff6c [eric] browser: refuse zero-LLM replay of send/submit skills, the live agent confirms those instead 2026-06-05 00:54:31 -07:00
ciregenz 9b6e45f365 [eric] browser: act-and-confirm, declare the expected change and verify it instead of assuming 2026-06-04 15:14:58 -07:00
ciregenz f0144690af [eric] browser: end the wait on the target or DOM-settle, not network-idle, so it stops waiting blind 2026-06-04 14:57:38 -07:00
ciregenz 2768d73a87 [eric] browser: log per-turn output tokens to measure terse-execution wins 2026-06-04 11:43:13 -07:00
ciregenz b77b29408f [eric] browser: cut per-action waste (skip discarded loop-hash, cache metrics dir) 2026-06-03 23:07:45 -07:00
ciregenz a562a5c3c3 [eric] browser: prune stale screenshots + capture downscaled JPEG to cut agent context cost 2026-06-03 22:17:05 -07:00
ciregenz bb422c0fe8 [eric] browser: treat a hung tab (repeated timeouts) as card-unavailable so it aborts fast, not a 20-min spin 2026-06-03 15:18:18 -07:00
ciregenz 9080f256c9 [eric] browser: log front-loaded perception reads so a read answered from them isn't flagged a ghost 2026-06-03 13:45:34 -07:00
ciregenz 61826d441a [eric] browser: BrowserRepeatFlow hands back capped per-item read data so batch-read actually delivers 2026-06-03 13:05:49 -07:00
ciregenz 21b225ded4 [eric] browser: nudge the agent toward the fast captured-API tier once per host 2026-06-03 00:50:23 -07:00
ciregenz 186c6cf72e [eric] browser: smart-wait returns when the page's network settles instead of a blind fixed sleep 2026-06-03 00:46:30 -07:00
ciregenz 7572462116 [eric] browser: intra-run batch replay (BrowserRepeatFlow) with verify gate, send-gate, network-route steps 2026-06-03 00:35:48 -07:00
ciregenz 4c0749d0db [eric] browser: subtle Remembered/Learned chip on the agent card so memory feels earned, zero clicks 2026-06-03 00:01:20 -07:00
ciregenz 3a601ba3ad [eric] browser: calm zero-click signals so the user feels the agent recall + learn site memory 2026-06-02 23:46:49 -07:00
ciregenz e6a1a5f2a5 [eric] browser: measure whether the playbook cuts exploration turns, flag when it doesn't 2026-06-02 23:19:47 -07:00
ciregenz f187ec7c6e [eric] browser: surface learned playbook in-chat + REST list/forget so site memory is visible and clearable 2026-06-02 23:16:22 -07:00
ciregenz fe082c492c [eric] browser: durable reflective per-site playbook (tier-2 memory, mem0-style distill+reconcile) 2026-06-02 23:11:47 -07:00
ciregenz 1208a2e746 [eric] browser: don't record a skill when the deliverable was gathered content, replay can't reproduce judgment 2026-06-02 21:20:20 -07:00
ciregenz 78f35aa9b4 [eric] browser: re-check for a replayable skill after the first nav lands on the right host 2026-06-02 19:06:21 -07:00
ciregenz 54d7e3c355 [eric] browser: key skills on the user's stable request, not the orchestrator's reformulation 2026-06-02 18:28:43 -07:00
ciregenz bcfbb8d24f [eric] browser: fail fast with an honest message when the card's webview is gone 2026-06-02 18:19:33 -07:00
ciregenz 34faca12ae [eric] browser: resolve the live card host for replay and report did-nothing runs honestly 2026-06-02 18:14:57 -07:00