Skip to content

feat: auto-repair reasoning_content para modelos Beam y bump a CLI 1.88.0 (v1.0.6) - #7

Merged
JoshRob297 merged 1 commit into
mainfrom
feat/reasoning-sanitizer-and-cli-update
Oct 26, 2026
Merged

JoshRob297 merged 1 commit into
mainfrom
feat/reasoning-sanitizer-and-cli-update

Conversation

@JoshRob297

Copy link
Copy Markdown
Owner

Resumen

  • Repara el error HTTP 400 Missing required parameter: messages[x].reasoning_content que impedia el uso de reflection/beam:free en conversaciones multi-turno o con agentes.
  • Inyecta de forma segura pensamientos sinteticos en sanitizeCommandCodeMessages unicamente cuando el modelo destino requiera razonamiento previo.
  • Actualiza la version de rastreo del CLI (CLI_VERSION) a 1.88.0.
  • Incluye tests unitarios especificos.

Verificaciones

  • bun test: 86 pass / 0 fail
  • bunx tsc --noEmit: limpio

…ion to 1.88.0 (v1.0.6)

- Support models requiring previous assistant reasoning_content in history (e.g. reflection/beam:free)
- Prevent HTTP 400 parameter errors in multi-turn assistant and agent workflows
- Update CLI_VERSION to 1.88.0 matching current npm release
- Unit tests added, 86 pass / 0 fail
@JoshRob297
JoshRob297 merged commit 49104df into main Oct 26, 2026
1 check passed
@JoshRob297
JoshRob297 deleted the feat/reasoning-sanitizer-and-cli-update branch October 26, 2026 02:43
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant