SurfSense/surfsense_local/backend/modules/agent
DESKTOP-RTLN3BA\$punk 666eed5974 fix(agent): take an optional id sent as 0 as left out
OpenAI's models fill every field a tool lists, so a render with no
document to continue and no template arrived with artifact_id 0 and
template_source_id 0. The tool asked for source 0, refused it as not
selected, and the model sent the same call until the turn ran out: in
the OpenRouter sweep no OpenAI model rendered a document, and two Grok
models hit the same refusal. Ids count from 1, so an optional *_id of 0
now names nothing.
2026-10-07 23:35:24 -07:00
..
agent_threads perf(agent): cut a step's nested inputs too, so a revision's edits stay out of its frames 2026-10-07 05:34:48 -07:00
data_analysis feat(agent): analyse spreadsheet sources with pandas scripts 2026-10-06 14:54:18 -07:00
model_endpoint fix(agent): keep a streamed line whole when it holds U+2028, U+2029 or U+0085 2026-10-07 02:14:55 -07:00
opencode_client fix(agent): keep a streamed line whole when it holds U+2028, U+2029 or U+0085 2026-10-07 02:14:55 -07:00
previews fix(ingest): retry what Docling cannot open yet; let it refuse a damaged deck 2026-10-06 22:42:35 -07:00
prompts Merge slice/inline-previews: a render's and a source's pages reach the model with the result 2026-10-04 22:38:42 -07:00
skills fix(agent): hold internal and standalone comments to the same sourcing rule 2026-10-06 19:24:59 -07:00
thread_folder fix(agent): commit before a sync waits for its thread's lock 2026-10-06 13:32:05 -05:00
tool_endpoint fix(agent): take an optional id sent as 0 as left out 2026-10-07 23:35:24 -07:00
__init__.py feat(agent): add the model endpoint opencode reaches every model through 2026-10-01 19:46:39 +02:00
dependencies.py feat(agent): let a chat thread be the agent's 2026-10-01 20:45:56 +02:00
engine_choice.py fix(chat): remember a picked mode for the model it was picked on 2026-10-07 16:03:12 -07:00
launch_key.py feat(agent): serve SurfSense's search to opencode as an MCP tool 2026-10-02 20:33:10 +02:00
model_call_route.py feat(local): call each remote model on its own route, and run the agent on a ChatGPT plan 2026-10-07 03:16:14 +05:30
model_reads_images.py feat(agent): show the agent its previews and source figures, and keep compaction out of the reply 2026-10-04 09:46:56 -07:00
model_window.py feat(local): call each remote model on its own route, and run the agent on a ChatGPT plan 2026-10-07 03:16:14 +05:30
opencode_config.py Merge dev into the file-agent gaps: per-model call routes and the agent on a ChatGPT plan 2026-10-06 20:24:11 -07:00
opencode_runtime.py feat(local): call each remote model on its own route, and run the agent on a ChatGPT plan 2026-10-07 03:16:14 +05:30