You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This commit was created on GitHub.com and signed with GitHub’s verified signature.
pi-commandcode-provider 0.7.1
Changes
Fix Gemini generate requests rejected by Vertex when tool parameters contain nullable type arrays. Preserve required fields and literal data without changing other models' schemas (#103, #99).
Refresh the Command Code CLI catalog to 1.56.0, including Qwen 3.8 Omni Flash image input, reasoning effort levels, a 131,072-token output limit, and reviewed display pricing (#102).
Correct npm-installed Oh My Pi plugin updates to use omp plugin install pi-commandcode-provider --force (#104).
Contributors
@JaimeGonzalezVallejo — reported, fixed, and live-validated Gemini nullable-schema handling.
@Newbie-troll — corrected the Oh My Pi update instructions.
Validation
Local release candidate: full npm test, real Pi against a mock API, typecheck, formatting, and git diff --check passed.
npm audit --audit-level=moderate: zero vulnerabilities.
Package dry-run: 29 files; no test artifacts or credentials included.
Included PRs passed GitHub CI, including Oh My Pi compatibility, security checks, and memory benchmarks.
Release PR #105 and post-merge main CI passed, including Oh My Pi compatibility.
Installed pi-commandcode-provider@0.7.1 from the public npm registry into an isolated Pi home and verified provider/model discovery, including Qwen 3.8 Omni Flash with image input, reasoning, and the 131,072-token output limit. Discovery used a harmless placeholder API key because Pi hides unconfigured providers; no authenticated model request was made.
Public registry version, latest dist-tag, and tarball integrity verified. No new live-auth completion test was run during this release.