All pull requests

Pull requests

• fix(catalog): an OSError on one skill is its error row, not a crash

  #417·napetrov opened Nov 6, 2026Contributor·

  ·
  5

• test(fuzz): atheris over the reply and frontmatter parsers; extract_json_object returns only objects

  #416·napetrov opened Nov 6, 2026Contributor·

  ·
  6

• ci(release): sign the sdist and wheel with Sigstore and attach SLSA build provenance

  #415·napetrov opened Nov 6, 2026Contributor·

  ·
  5

• feat(catalog): corpus mode for catalog scan (#413 A3, A4, A6, A7)

  #414·napetrov opened Nov 6, 2026Contributor·

  ·
  14

• fix(mcp): an mcp: source's ;id= wins over the registry in every command

  #411·vishalhabib99 opened Nov 5, 2026Contributor·6 tasks done·

  ·
  1

  5

• feat(eval): warn on self-judging on every judge path that lacked it

  #408·napetrov opened Nov 5, 2026Contributor·

  ·
  28

• fix(novelty): write the partial routing result when --max-usd stops the run

  #406·napetrov opened Nov 5, 2026Contributor·

  ·
  18

• fix(novelty): keep a skill whose only spec error is an over-long description in the routing field

  #405·napetrov opened Nov 5, 2026Contributor·

  ·
  17

• fix(novelty): check routing --out before calls, clearer YAML errors, real output estimate, noise shape

  #403·napetrov opened Nov 4, 2026Contributor·

  ·
  30

• feat(task-formats): judge custom scorers, read hf_dataset and code promptfoo configs, name Harbor suites

  #402·napetrov opened Nov 4, 2026Contributor·

  ·
  66

• test(fidelity): golden of the request a docker-claude arm sends, checked in CI

  #400·napetrov opened Nov 4, 2026Contributor·

  ·
  56

• feat(eval): eval tasks generate --from graph: enumeration tasks from a SCIP index

  cli-interface-change
  #399·napetrov opened Nov 4, 2026Contributor·

  ·
  48

• fix(repo): Bash reads #374 left: substitutions, activate, case words, time, shell functions

  #397·antonwolfy opened Nov 4, 2026Contributor·6 tasks done·

  ·
  43

  

• refactor(cli): move skill, catalog, trace and data under their roots (#94 M2)

  cli-interface-change
  #396·napetrov opened Nov 4, 2026Contributor·

  ·
  8

• feat(repo): repo compare --metric judge:recall_judged reads judged recall

  #395·antonwolfy opened Nov 4, 2026Contributor·6 tasks done·

  ·
  1

  23

  

• fix(routing): name what the field leaves out; add --keep-copies-under and all-vs-all estimate

  cli-interface-change
  #394·napetrov opened Nov 4, 2026Contributor·

  ·
  38

• fix(cli): raise InputError for input refusals; run failures exit 1 (part of #275)

  #393·napetrov opened Nov 4, 2026Contributor·

  ·
  19

• feat(repo): repo compare flags replay tasks a guidance revision may have seen

  #392·antonwolfy opened Nov 4, 2026Contributor·6 tasks done·

  ·
  1

  35

  

• fix(config): nested-repo ignore rules and per-file eval suites (follow-up to #323)

  #388·napetrov opened Nov 4, 2026Contributor·

  ·
  30

• ci: build and test conda packages and wheels

  enhancement
  #378·jharlow-intel opened Nov 3, 2026Contributor·5 of 6 tasks·

  ·
  52