Skip to content

v0.3.291

Latest

Choose a tag to compare

@github-actions github-actions released this 17 Nov 22:04
8012060

0.3.291 (2026-11-17)

Features

  • count samples of a finished log as failures with an empty-attempt run and inspect log zero (#6579) (7e056a2)
  • eval-set selection workers can run their tasks with empty_attempt() for zero_samples() (#6605) (2dada81)
  • exclude samples of a finished log from its metrics, from Python or inspect log exclude (#6543) (dcb464f)
  • hf: option to ignore a model repository's generation config, and record template and generation settings in the log (#6610) (33308b0)
  • inspect log merge-shards merges a sharded run's logs from the command line (#6581) (5387a26)
  • openai: support GPT-6.2 Astra and GPT-6.2 Sol (#6614) (25de466)
  • sentinel: ask a person about a call when a sentinel sidecar escalates it (#6519) (b181668)
  • sentinel: ask a person about a model's whole turn when a sentinel sidecar escalates it (#6596) (d30a177)
  • sentinel: ask a person about a request when a sentinel sidecar escalates it (#6617) (bcc1d85)
  • sentinel: human() can ask a person about a request before the model is sent it (#6615) (b1f6078)
  • sentinel: run sentinels for bridged agents (#5765) (c7f96da)
  • update Fireworks and Together model info (#6426) (ca0ab01)

Bug Fixes

  • a bridged host tool answered with the agent's own timeout is no longer linked as its result (#6523) (d59b5e2)
  • ACP clients show the agent as still working after a turn ends on a limit or error (#6553) (8c2e11f)
  • an MCP tool called while its local server is still connecting no longer runs ahead of the handshake or fails with it (#6552) (cf2e2d7)
  • Anthropic usage undercounted when a request compacts its context (#6390) (998f93a)
  • exclude sentinel, approver and reviewer spend from cost limits; add suspend_cost_limit() and suspend_limits() (#6517) (bf036e4)
  • exec_remote start whose command closes stdin before reading its input always fails with "Connection lost" instead of sometimes starting with the input dropped (#6540) (e5dee1c)
  • Gemini computer-use calls lose Gemini's intent field and exact coordinates on replay (#6589) (3f34c35)
  • harden checkpoint snapshot IDs against option injection (#6580) (b01cb66)
  • install_skills() inside a sandbox identity scope fails for a skill with scripts when the image's working directory is closed to the scope's user (#6583) (ce19dd3)
  • model completions in logs from before 0.3.123 read back as text instead of an attachment:// reference (#6587) (3276a52)
  • model text that begins with attachment:// or tc:// reads back from the log as written (#6584) (ab2122c)
  • OpenAI native compaction crashes when a user message contains an image or file (#6436) (92fa034)
  • reduce probability of Docker sandbox project name collisions (#6575) (386b487)
  • remote MCP server credentials are no longer written to eval logs (#6492) (8582920)
  • replay pairs each tool call with its own approvals and sentinel checks when a call id repeats (#6370) (db7be4e)
  • S3 downloads keep extra chunks in memory on a busy machine (#6586) (f442184)
  • samples_df no longer uses pandas' deprecated drop_duplicates(inplace=True) (#6383) (a3b8090)
  • sandbox calls inside a sandbox identity scope default to the scope's working directory (#6539) (a2e99b2)
  • tools with an Enum, date or Path parameter default failing every model call (#6458) (f68c610)