All pull requests

Pull requests

• Add FireHydrant investigations toolset and server-side poller

  #2721·arikalon1 opened Nov 8, 2026Collaborator·

  ·
  14

• ROB-1511 Run the credential-blocked evals against self-hosted backends

  #2720·naomi-robusta opened Nov 8, 2026Collaborator·

  ·
  105

• Weekly Benchmark Results 2026-11-08_03-06

  #2719·github-actions[bot] opened Nov 8, 2026Bot·

  ·
  2

• docs: add OpenLLM to the AI providers

  #2718·apewithllm opened Nov 7, 2026·

  ·
  7

• Let the model read files bundled with a skill

  #2717·officialasishkumar opened Nov 7, 2026Contributor·

  ·
  1

  10

• Restore the terminal when holmes exits during a live display

  #2716·officialasishkumar opened Nov 7, 2026Contributor·

  ·
  1

  9

• Don't send MCP results to the model twice when the copies differ in form

  #2715·officialasishkumar opened Nov 7, 2026Contributor·

  ·
  1

  8

• Add OpenUI visualization evals

  #2712·avshyz opened Nov 6, 2026Contributor·4 tasks·

  ·
  19

• ROB-1761 Fail the benchmark run when no test ran

  #2708·RoiGlinik opened Nov 5, 2026Collaborator·

  ·
  7

• ROB-1761 Treat the string "null" as null for optional tool params

  #2707·RoiGlinik opened Nov 5, 2026Collaborator·

  ·
  3

• ROB-1761 Benchmark haiku-5.5-medium instead of haiku-4.5

  #2706·RoiGlinik opened Nov 4, 2026Collaborator·

  ·
  7

• Weekly Benchmark Results 2026-11-04_15-04

  #2705·github-actions[bot] opened Nov 4, 2026Bot·

  ·
  2

• ROB-1671 Run approval-gated tools remotely after a verified human approval

  #2696·naomi-robusta opened Nov 4, 2026Collaborator·

  ·
  11

• FRO-620 docs: strict build passes, fence errors name the key's line, faster fence checks

  #2693·ezra-robusta opened Nov 2, 2026Collaborator·

  ·
  19

• FRO-630 docs: name the Holmes service account by release on the AWS page

  #2692·ezra-robusta opened Nov 2, 2026Collaborator·

  ·
  2

• fix(tools): drop headers with shell metacharacters instead of raising error

  #2691·cf3901646 opened Nov 2, 2026·

  ·
  1

  3

• Add TOOL_MEMORY_LIMIT_PERCENT and TOOL_MAX_PARALLEL_CALLS

  #2687·xval2307 opened Oct 30, 2026·

  ·
  7

• docs: add a Redis (MCP) page to the built-in toolsets

  #2686·jhwanseok opened Oct 30, 2026·

  ·
  14

• Fix "build-in" typo in toolset CLI help and other misspellings

  #2684·sarankumar1325 opened Oct 29, 2026·

  ·
  4

• Add a holmesgpt.toolset.status gauge for toolset health

  #2681·vivek-raj1 opened Oct 29, 2026·

  ·
  1

  15

• fix(otel): Report per-call token usage and cache tokens on LLM spans

  #2672·Akshay-Akz2004 opened Oct 27, 2026·

  ·
  1

  10

• Move Jira source search to /rest/api/3/search/jql

  #2671·torrresagus opened Oct 27, 2026·

  ·
  3

• Normalise empty tool-call arguments to an empty JSON object

  #2670·xubenqiang opened Oct 26, 2026·

  ·
  1

  9

• Remove the calendar-year cap from certifi

  #2654·mday-io opened Oct 23, 2026·3 of 4 tasks·

  ·
  3

• fix(chat): return an empty analysis when the model produces no final text

  #2652·cvas22 opened Oct 22, 2026·

  ·
  1

  22