Tags: ifixai-ai/iFixAi
Tags
feat(cli): align skill + plugin onto the guided ifixai run engine (#57) * feat(cli): align skill + plugin onto the guided `ifixai run` engine Scaffold /ifixai-skill into coding agents (ifixai install), converge the Claude marketplace plugin onto ifixai run, add run-health validity banners, and remove the retired ifixai-diagnose orchestrator backend. Docs + README updated for the universal install path. * feat(plugin): add Codex plugin, show display name as iFixAi on both hosts Commit the untracked Codex manifest and marketplace so the port ships, keep manifests at 0.1.1 to match the Claude plugin, and document both Codex run paths in the README. * fix(docs): replace dead IFIXAI_MAX_LIVE_CALLS with real concurrency knob SKILL.md and playbook.md now use IFIXAI_CONCURRENCY=1 / --no-parallel for sequential runs, and bridge.py drops the deleted orchestrator's --mode api reference. * docs(readme): clarify plugin vs skill in the run sections Reorganize by mechanism: plugin (Claude Code + Codex, preferred) and skill (/ifixai-skill, every agent incl. both hosts). Fix the comparison table and intro that implied the plugin works in all agents. * fix: apply pre-merge review findings Harden the operator playbook + SKILL against splicing untrusted repo values into a shell and make the --dry-run consent step mandatory; revert two getattr fallbacks to direct attribute access; make bootstrap wording host-neutral for Codex as well as Claude. * fix(reports): suffix report filenames with the run nonce Each run's summary/full/json now share a unique run-nonce suffix, so a later run with a different provider or fixture no longer leaves a stale report that reads as current. Operator docs point at the exact printed paths.
feat: iFixAi 3.0.2 — Claude Code plugin + PyPI packaging (#47) * feat: iFixAi 3.0.2 — Claude Code plugin + PyPI packaging - Installable Claude Code plugin (skill, hooks, manifest) + multi-provider SUT/judge engine and interactive scorecard artifact - pip-install ready: ifixai-diagnose entry point, ifixai.plugin package-data, version 3.0.2, tag-triggered PyPI release workflow - Refreshed README + Diátaxis docs/ * plugin: model preflight and run-health gates for live runs A per-model preflight aborts a live run on a bad model id before billing; a measurement-failure banner suppresses the grade when the agent or judge was unreachable; SUT transport errors now route to unscorable (base.py, b19/b20/b22) so a dead model scores INCONCLUSIVE rather than a false FAIL; adds a judge-health note and drops the dead b07 structural path. * fix: gate platform-refusal to bridge mode, restore B01/P01 fail-closed, guard judge-timeout parse Stops a legitimate live-API refusal that cites a usage policy from being dropped as unscorable (the wall check now runs only for bridge-routed SUT replies), reverts the B01/P01 capability-missing carve-out so an ungoverned target fails closed and caps at D per docs/scoring.md, and parses IFIXAI_JUDGE_TIMEOUT defensively so a bad value no longer crashes import. Also fixes the bridge str|None mypy error, drops dead _checkpoint_path, and dedups the grade constants into grading.py. * fix: make plugin Windows-compatible (bootstrap.py + PS1 shim + exec-form hook) Move provisioning logic into a cross-platform bootstrap.py (bin/ vs Scripts/, python3/python fallback). Add bootstrap.ps1 for native Windows PowerShell. Switch the SessionStart hook to exec form (python3 + python fallback entries) so no sh is needed. Update SKILL.md with Windows paths and shim invocations. * fix: Windows-compat follow-ups (UTF-8 console, LF gitattributes, dedup bootstrap hook) * Set plugin author and marketplace owner to iFixAi * docs: show unique-cloners chart in README (parity with main)
feat: New benchmarks (#36) * chore: temp commit p01 * chore: black formatter * chore: P08 benchmark * chore: remove changelog * chore P13 implementation * chore P19 implementation * chore P22 implementation * chore: P27 implementation * chore: P32 implementation * chore: C02 implementation * chore: C05 implementation * chore: C1 implementation * chore: S02 implementation * chore: X04 implementation * chore: X11 implementation * chore: replace private judge-internals reach with public pipeline accessors * feat: category-filtered runs + shared evidence builder * chore: remove test script
feat: improve b19–b23 unpredictability inspections (#34) Update runners, rubrics, references, and definitions for the context-accuracy (b19), instruction-adherence (b20), cross-turn-objective (b21), decision-reproducibility (b22), and policy-version-trace (b23) inspections. Add b21 step-1 rubric, a concurrency tweak, scorecard update, and docs refresh. Ignore locally-regenerated run artifacts. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
chore: fabrication benchmark investigation (#30) * chore: B10 improvement * chore: B14 improvement * chore: B15 improvement * chore: B17 improvement * chore: B17-updated improvement * chore: B18 improvement * chore: B24 improvement * chore: B25 improvement + B24 hotfix + B18 speed optimization * chore: B26 improvement * chore: B27 improvement * chore: B29 improvement * chore: B31 improvement + fixture example updates * chore: B32 improvement * chore: B17 speed fix + B31 correct fixture mapping + better result detectiong * chore: B01 improvement * chore: B02 improvement * chore: B03 improvement * chore B04 improvement * chore B05 improvement + better fixtures * chore: B06 update * feat: multi-benchmark selection via repeatable * chore: CI fix * chore: remove unused imports
PreviousNext